Author: Stackai

  • PSG Grant for AI Automation in Singapore: The Complete 2026 Guide for SMEs

    PSG Grant for AI Automation in Singapore: The Complete 2026 Guide for SMEs

    The Productivity Solutions Grant (PSG) is a Singapore government grant, administered by Enterprise Singapore, that funds up to 50% of the cost of pre-approved digital and automation solutions for local SMEs. For Singapore business owners weighing the cost of AI automation — chatbots, invoice processing, HR onboarding workflows, or CRM automation — PSG is usually the fastest and simplest grant to tap, because it works off a pre-approved vendor list rather than requiring a full project proposal like the Enterprise Development Grant (EDG).

    This guide covers what PSG funds, who qualifies, what it costs after funding, how to apply step by step, how it compares to EDG, and what’s changing when Enterprise Singapore’s new EDGE grant framework launches later this year.

    “PSG applications must be submitted before any contract is signed or payment made — there is no retroactive funding.”

    What Is the Productivity Solutions Grant (PSG)?

    PSG is a Singapore government co-funding scheme that helps SMEs adopt pre-scoped, off-the-shelf digital solutions — including AI-powered automation tools — without having to write a custom project proposal. Instead of assessing a bespoke project (as EDG does), PSG works from the GoBusiness Tech Depot: a curated list of solutions from approved vendors, each pre-assigned a maximum fundable cost. This makes PSG the more accessible entry point for SMEs that want to automate a specific, well-defined function — such as WhatsApp customer service, invoice processing, or appointment scheduling — rather than a full agentic AI system spanning multiple departments.

    Is Your Business Eligible for PSG?

    Enterprise Singapore sets five core eligibility criteria for PSG applicants:

    Criterion Requirement
    Registration Registered and operating in Singapore
    Local equity At least 30% held by Singaporean citizens and/or PRs, direct or indirect
    Company size Group annual turnover under S$100 million OR group employment under 200 employees
    Solution use The IT solution must be deployed and used in Singapore
    Sequencing No contract signed and no payment made to the vendor before the PSG application is submitted

    One detail trips up more applicants than any other: PSG applications must be submitted before any contract is signed or payment made. If a business engages a vendor first and applies for PSG afterward, Enterprise Singapore will reject the application automatically — there is no retroactive funding.

    What Can PSG Fund for AI Automation?

    PSG funds a wide range of pre-approved categories relevant to Singapore SMEs pursuing automation, including customer management and CRM systems, HR and payroll automation, accounting and invoicing automation, inventory and procurement systems, and increasingly, AI-powered chatbot and workflow automation tools listed under the digital solutions categories. The table below summarizes the funding mechanics.

    Item Detail Notes
    Funding level Up to 50% of qualifying costs For pre-approved solutions on the PSG Tech Depot
    Annual cap S$30,000 per company, per financial year Across all PSG applications combined
    Eligible spend Pre-approved IT solutions and equipment Must be selected from the GoBusiness Tech Depot listing
    Application timing Before signing any vendor contract or making payment Applying after a contract is signed leads to automatic rejection
    Processing time Typically 4-6 weeks Via the Business Grants Portal (BGP)

    In practice, this means a Singapore SME automating, for example, WhatsApp-based customer enquiries and appointment booking could have up to half the solution cost covered by PSG, up to the S$30,000 annual cap — with the remaining cost potentially eligible for the Enterprise Innovation Scheme’s 400% tax deduction on top.

    How to Apply for PSG: Step-by-Step Process

    • Identify a pre-approved solution on the GoBusiness Tech Depot that matches your automation need, or work with a vendor who can help scope one.
    • Obtain a formal quote from the vendor — do not sign a contract or make any payment at this stage.
    • Log in to the Business Grants Portal (businessgrants.gov.sg) using CorpPass.
    • Submit the PSG application with the vendor quote and required company information attached.
    • Wait for approval — typically 4 to 6 weeks — and receive a Letter of Offer.
    • Accept the Letter of Offer within the stated deadline, then engage the vendor and begin deployment.
    • After deployment, file a claim through the BGP with invoices, proof of payment, and deployment confirmation.
    • Receive the 50% reimbursement into your nominated company bank account.

    Not sure if PSG or EDG fits your project?

    We scope every automation project against the grant it actually qualifies for — before you apply, not after.

    Book a Free Call

    PSG vs. EDG: Which Grant Fits Your AI Automation Project?

    SMEs evaluating AI automation often ask whether to apply for PSG or EDG — the two are not interchangeable, and the right choice depends on project scope:

    • Choose PSG when the automation need maps to a pre-approved, off-the-shelf solution — a single-function tool like a chatbot platform, invoicing system, or scheduling automation. Faster approval, simpler application, lower funding cap.
    • Choose EDG when the project is a custom, multi-function agentic AI system built around your specific processes — EDG funds up to 50% (and in some cases higher) of bespoke consultancy and development costs with a higher overall ceiling, but requires a fuller project proposal.
    • Some Stack AI engagements qualify for a blended approach — a pre-approved component under PSG alongside a custom-scoped component under EDG — though this needs to be structured carefully with Enterprise Singapore’s guidelines in mind.

    Important: The EDGE Grant Is Coming in Late 2026

    Enterprise Singapore has announced a new consolidated grant framework called EDGE, expected to launch in the second half of 2026, which will merge EDG, PSG, and the Market Readiness Assistance (MRA) grant into a single application through the Business Grants Portal. Until EDGE launches, PSG, EDG, and MRA all remain fully accessible under their current rules — businesses with a project ready to go do not need to wait. SMEs currently planning an AI automation project should treat this as a reason to move sooner rather than later: applying under the existing PSG rules now avoids any uncertainty around how eligibility, caps, or the application process may change once EDGE takes effect. Stack AI is tracking the EDGE rollout closely and will update grant guidance as details firm up.

    Frequently Asked Questions

    Can PSG fund an AI chatbot or WhatsApp automation system?

    Yes, if the solution is listed on the GoBusiness Tech Depot or offered by a vendor whose relevant solution has been pre-approved under a PSG category such as customer management or digital solutions. This covers many AI-powered chatbot and customer service automation tools.

    How much can my business actually receive from PSG?

    Up to 50% of the qualifying solution cost, capped at S$30,000 per company per financial year across all PSG applications combined. The exact per-solution cap depends on the maximum fundable amount listed for that specific solution on the Tech Depot.

    Can I apply for PSG after I’ve already started using a solution?

    No. PSG requires the application to be submitted, and typically approved, before any contract is signed or payment is made to the vendor. Starting first and applying for reimbursement afterward is not permitted.

    Can PSG and EDG be used on the same project?

    They generally fund different cost categories — PSG for pre-approved off-the-shelf solutions, EDG for custom consultancy and development — so a single large automation project can sometimes combine both, but this requires careful scoping with Enterprise Singapore’s current guidelines and is best discussed directly with a grant-experienced vendor.

    Will PSG still exist once the EDGE grant launches?

    Enterprise Singapore has stated that EDG, PSG, and MRA will be consolidated into EDGE when it launches in the second half of 2026. Businesses should expect PSG as a standalone grant to be phased out at that point, replaced by a single EDGE application covering the same funding areas.

    Get Grant-Funded AI Automation Done Right

    Stack AI helps Singapore SMEs scope, apply for, and deploy AI automation projects that qualify for PSG, EDG, TSS, and the Enterprise Innovation Scheme’s 400% tax deduction — so more of the automation budget is government-funded, not out of pocket. Book a free call to find out which grant fits your project.

  • AI Automation for GP Clinics in Singapore: The 2026 Playbook

    AI Automation for GP Clinics in Singapore: The 2026 Playbook

    Singapore’s GP and medical clinics run on tight margins and even tighter front-desk headcount. A solo practice or small group clinic often has one or two staff juggling phone calls, WhatsApp enquiries, walk-ins, and appointment changes at the same time, while patients increasingly expect the instant, 24/7 responsiveness they already get from e-commerce and food delivery apps. This playbook covers where AI automation delivers the fastest return for a Singapore clinic, the data-privacy and clinical-safety guardrails that apply, how much of the cost is grant-fundable, and what a typical Stack AI clinic engagement looks like.

    “Patients increasingly expect the instant, 24/7 responsiveness they already get from e-commerce and food delivery apps — not a phone line that’s busy until 9am.”

    Where Clinics Lose the Most Time

    Front-desk teams at Singapore clinics tend to lose time to the same handful of repetitive, high-volume tasks:

    Pain Point Typical Impact
    Repetitive WhatsApp/phone enquiries (hours, doctor availability, walk-in wait, directions) Front-desk staff interrupted dozens of times a day; slower service for patients physically in the clinic
    Appointment booking and rescheduling handled manually Double-bookings, long hold times, no booking capacity outside clinic hours
    No-shows Empty slots, lost revenue, no consistent reminder cadence
    Post-visit follow-ups (medication reminders, review bookings, chronic-disease check-ins) Inconsistent follow-through and missed repeat-visit revenue
    Claims and billing admin (insurance, Medisave, corporate panels) Manual data entry and delayed reimbursement

    What Stack AI Can Automate for a Clinic

    Stack AI automation is designed to sit alongside a clinic’s existing Clinic Management System (CMS), including systems on MOH’s CMS Tiering Framework for Primary Care, rather than replace it. Automation connects to the existing patient records and booking system via API, so a clinic does not need to switch CMS vendors to benefit. Typical automation covers:

    A 24/7 WhatsApp or website chat assistant that answers FAQs, checks appointment availability, and books or reschedules directly into the clinic’s existing calendar or CMS.

    Automated appointment reminders (WhatsApp/SMS) with two-way confirm-or-cancel replies, which is the single fastest way most clinics cut no-shows.

    Post-visit follow-up sequences, covering medication reminders, review-visit scheduling, and chronic-condition check-ins, that run automatically instead of relying on staff to remember.

    Pre-visit intake automation, collecting patient forms and history digitally before the patient arrives.

    Claims and billing admin automation, capturing structured data and routing it to reduce manual entry for insurance and corporate-panel claims.

    Internal operations support, including automated end-of-day appointment summaries, staff scheduling reminders, and stock reorder alerts for consumables.

    Data Privacy and Clinical Safety: The Non-Negotiables

    Healthcare automation carries a different risk profile than a retail or F&B chatbot. Every clinic automation project Stack AI scopes is built around a few non-negotiables, before a single workflow is written:

    PDPA compliance for patient data handling, with clear consent capture at first contact.

    No AI-generated clinical advice or diagnosis. The assistant handles scheduling, logistics, and FAQs, and hands off to clinical staff for anything involving symptoms, results, or treatment.

    Human-in-the-loop escalation for any message a patient flags as urgent, or that the AI cannot confidently resolve.

    Alignment with MOH and HSA’s AI in Healthcare Guidelines (AIHGIe 2.0, refreshed March 2026), which set expectations for transparency, risk mitigation, and defined responsibilities between developers, deploying clinics, and clinicians.

    Funding Your Clinic’s Automation Project

    Most clinic automation projects Stack AI builds, including WhatsApp assistants, reminder systems, and intake and billing automation, qualify as pre-approved or custom digital solutions, making them fundable the same way as other Singapore SME automation projects:

    Grant What It Covers Funding
    PSG (Productivity Solutions Grant) Pre-approved, off-the-shelf automation tools: chatbot platforms, scheduling, invoicing Up to 50%, capped at S$30,000 per company per financial year
    EDG (Enterprise Development Grant) Custom, multi-function automation built around the clinic’s specific workflows Up to 50% of qualifying consultancy and development cost
    Enterprise Innovation Scheme Tax deduction on qualifying automation spend, on top of grant funding 400% tax deduction

    Ready to see what’s fundable?

    We grant-pre-check every clinic automation project before a contract is signed.

    Book a Free Call

    How a Clinic Automation Project Works

    1. Discovery call: walk through the current booking flow, CMS/software stack, and where staff time is actually going.

    2. Scope the highest-impact automation first, usually WhatsApp booking plus reminders, since it deploys fastest and shows ROI within weeks.

    3. Grant pre-check: confirm whether the project maps to a PSG pre-approved category or needs EDG scoping, before any contract is signed or payment made.

    4. Build and connect: integrate with the existing CMS/calendar via API, and configure escalation rules and PDPA-compliant consent flows.

    5. Staff walkthrough and soft launch: the front-desk team is trained, and the assistant handles a subset of enquiries alongside staff.

    6. Full rollout and monitoring: track no-show rate, response time, and staff hours saved, then tune the assistant based on real patient enquiries.

    Frequently Asked Questions

    Can an AI chatbot actually book appointments into our existing clinic software?

    Yes, provided the CMS or booking system has an API or supports calendar integration; most modern Singapore clinic systems do. Stack AI connects the WhatsApp or chat assistant directly to the existing calendar rather than requiring a new booking system.

    Will patients be talking to AI instead of clinic staff?

    For logistics such as booking, rescheduling, reminders, and FAQs, yes, and most patients prefer the speed. For anything involving symptoms, results, or clinical judgment, the assistant is scoped to hand off to staff rather than answer.

    Is this compliant with PDPA and MOH’s AI guidelines?

    Every clinic project is scoped with PDPA-compliant consent capture and aligned with MOH/HSA’s AIHGIe 2.0 guidelines. Stack AI does not build systems that provide diagnosis or clinical advice.

    How much does clinic automation typically cost after grant funding?

    It depends on scope, but a WhatsApp booking and reminder system typically qualifies under PSG’s pre-approved categories, meaning up to 50% of the solution cost (capped at S$30,000 per year) is covered, before the Enterprise Innovation Scheme’s tax deduction is applied on top.

    How long does it take to go live?

    A first-phase WhatsApp booking and reminder automation typically takes a few weeks to configure and connect once a solution is confirmed. Full CMS integrations and multi-workflow builds take longer, depending on scope.

    Automate Your Clinic’s Front Desk, Without the Compliance Risk

    Stack AI helps Singapore GP and medical clinics automate front-desk admin, including booking, reminders, follow-ups, and billing support, while staying inside PDPA and MOH’s AI healthcare guidelines, with up to 50% of the cost grant-fundable. Book a free call to scope your clinic’s automation project.

  • TSS Grant for AI Automation in Singapore: The Complete 2026 Guide for Social Service Agencies

    TSS Grant for AI Automation in Singapore: The Complete 2026 Guide for Social Service Agencies

    What Is the TSS Grant?

    The Transformation Sustainability Scheme (TSS) is a National Council of Social Service (NCSS) grant that funds social service agencies (SSAs) in Singapore to strengthen how they operate — including funding for digital projects and AI-powered automation. TSS replaced the earlier Tech-and-GO! scheme and now covers a broader range of organisational development needs, from leadership consultancy to bespoke chatbot builds.

    For SSAs considering AI automation — intake chatbots, case management workflows, donor and volunteer coordination, appointment scheduling, or back-office finance and HR automation — TSS is the primary funding route, and at up to 80% co-funding it is one of the most generous grants available to any organisation type in Singapore, government or private sector.

    This guide covers eligibility, what TSS actually pays for, how it compares to the grants SMEs typically use (EDG, PSG, EIS), and the exact steps to apply.

    “At up to 80% co-funding, TSS is one of the most generous grants available to any organisation type in Singapore — government or private sector.”

    Who Is Eligible

    • Your organisation is an NCSS member agency, or is funded by the Ministry of Social and Family Development (MSF).
    • You hold a valid Organisational Health Framework for Social Services (OHFSS) assessment — or are willing to complete one before applying.
    • For non pre-scoped funding (Parts A and B, covered below), your agency has completed the Organisational Health Diagnostics (OHDS) process.

    If your organisation is not an NCSS member or MSF-funded SSA — for example, a private clinic, tuition centre, or F&B business — TSS does not apply to you. Those organisations should instead look at EDG, PSG, or EIS; see the comparison table below.

    What TSS Actually Funds

    TSS is structured into three parts, each with a different funding cap and use case. Up to 80% co-funding applies across all three, over a project term of up to 3 years.

    Component What it covers Funding cap
    Part A — Consultancy & project implementation support Engaging consultants (e.g. compensation & benefits review, volunteer management framework) or funding project headcount to lead a transformation project. Up to $250,000 (headcount support capped at $200,000)
    Part B — Bespoke digitalisation projects Custom-built digital systems: web portals for service delivery, AI chatbots for automated responses, therapeutic robotics, and other purpose-built AI automation. Up to $150,000
    Part C — Pre-scoped & green-laned solutions Off-the-shelf systems from NCSS’s approved vendor list: HR systems, accounting systems, event booking platforms, CRM systems, API integrations. Up to $150,000 total, capped at $40,000 per solution

    Part B is the most relevant component for custom AI automation work — NCSS explicitly names “AI chatbots for automated responses” as a funded example, alongside therapeutic robotics and custom portals. Part C is the faster route if a pre-scoped or Green Lane solution already covers your need; check NCSS’s published solutions list before scoping a custom build.

    TSS vs. the Grants SMEs Normally Use

    Stack AI works across both SME and social-service clients, and the two groups draw on different grants. If you are unsure which applies to you, use this table:

    Grant Who it’s for Typical co-funding Best for
    TSS NCSS-member or MSF-funded social service agencies Up to 80% AI chatbots, case management automation, custom digital systems, org-wide transformation projects
    EDG Singapore-registered SMEs (30% local shareholding) Up to 50% Business/process transformation projects, including AI workflow automation
    PSG Singapore-registered SMEs Up to 50% Pre-approved, off-the-shelf digital solutions (faster, lighter-touch than EDG)
    EIS Any Singapore-registered business investing in innovation/automation 400% tax deduction (not co-funding) Businesses that prefer a tax deduction over a cash grant, or that exceed EDG/PSG cofunding caps

    A note on eligibility overlap: some social enterprises and charity-run social businesses may qualify for both TSS and SME-track grants depending on their registration status and funding source — worth confirming with NCSS and Enterprise Singapore directly rather than assuming one or the other applies.

    What This Looks Like in Practice

    Examples of AI automation projects social service agencies have funded (fully or partly) through schemes like TSS Part B and Part C:

    • An intake chatbot that screens and triages new client referrals before a caseworker is assigned, cutting initial response time from days to minutes.
    • Automated case-note summarisation that drafts structured notes from caseworker voice memos or free-text input, reducing admin time per case.
    • Volunteer and donor CRM automation that syncs sign-ups, sends reminders, and flags lapsed donors for follow-up without manual list management.
    • Appointment and programme-booking automation for community centres and eldercare/childcare services, replacing phone-based scheduling.
    • Finance and grant-reporting automation that pulls data from multiple systems into the reports NCSS and MSF require, reducing month-end close time.

    Building a Part B proposal?

    Stack AI scopes bespoke AI automation builds the way NCSS expects — vendor quotation included.

    Book a Free Call

    How to Apply: Step by Step

    • Confirm eligibility — verify NCSS membership or MSF funding status, and check whether your OHFSS assessment is current.
    • Complete Organisational Health Diagnostics (OHDS) if you haven’t already — required before applying for Part A or Part B funding.
    • Scope the project and decide which part applies — Part A (consultancy/headcount), Part B (bespoke digital/AI build), or Part C (pre-scoped/Green Lane solution).
    • Get a vendor quotation — for a bespoke AI automation build, this means a scoped proposal from an AI automation vendor (this is where Stack AI comes in for Part B projects).
    • Submit your application through the OurSG Grants (OSG) portal, with your OHFSS report and vendor quotation attached.
    • On approval, sign the vendor agreement (and submit proof of employment if claiming headcount support under Part A).
    • Implement the project, then submit your completion report, updated OHFSS report, and (for Part A) Employee Engagement Survey and Board Assessment reports to close out the grant.

    Processing timelines and exact documentation requirements are set by NCSS and can change — always confirm current requirements at ncss.gov.sg or by emailing sector_capability@ncss.gov.sg before submitting.

    Frequently Asked Questions

    Is TSS only for large, established charities?

    No. Eligibility is based on NCSS membership or MSF funding status, not organisation size. Smaller SSAs are still eligible, and Part C’s pre-scoped solutions in particular are designed to be accessible without a large internal project team.

    Can TSS fund an AI automation project run by an external vendor like Stack AI?

    Yes — Part B (bespoke digitalisation) and Part C (pre-scoped/Green Lane solutions) are both designed to fund vendor-built or vendor-supplied systems, provided you submit a vendor quotation as part of your application and the vendor agreement is signed after approval.

    What’s the difference between Part B and Part C for an AI chatbot project?

    Part C is for solutions already on NCSS’s pre-scoped or Green Lane list — faster to apply for, capped at $40,000 per solution. Part B is for a bespoke build tailored to your workflows — higher cap ($150,000) but requires the OHDS diagnostic step first. Check the pre-scoped solutions list before assuming you need a custom build.

    Does TSS cover ongoing costs, or only the initial build?

    TSS funding is project-based and tied to implementation — it is not designed as a recurring operating subsidy. Ongoing software subscription or maintenance costs should be discussed with NCSS or your assigned grant officer as part of project scoping, since treatment can vary by project.

    How does TSS relate to the EDGE grant merging EDG, PSG, and MRA for SMEs?

    They are separate tracks. EDGE (still not launched as of this guide’s writing) consolidates SME-facing grants under Enterprise Singapore. TSS is an NCSS/MSF scheme specifically for the social service sector and is not part of that consolidation.

    Why Work With Stack AI

    Stack AI builds custom AI automation systems — chatbots, case workflow automation, CRM and reporting integrations — for Singapore organisations, and structures projects to align with grant requirements from the scoping stage rather than after the fact. For social service agencies, that means a Part B or Part C-ready proposal, a vendor quotation formatted the way NCSS expects, and implementation support through to your completion report.

    If your agency is evaluating an AI automation project under TSS, get in touch to scope what’s fundable before you apply.

  • AI Automation for Tuition Centres in Singapore: The 2026 Playbook

    AI Automation for Tuition Centres in Singapore: The 2026 Playbook

    Singapore households spent an estimated S$1.8 billion on private tuition in 2024 — up nearly 30% since 2018 — and roughly 7 in 10 students attend some form of tuition. There are now over 1,000 tuition and enrichment centres competing for those parents. The market is large, but it is also crowded, and the centres that grow are rarely the ones with the best marketing budget. They are the ones that reply to a trial-class enquiry in two minutes instead of two days, collect fees without awkward chasing, and keep parents feeling informed enough that they never think about switching.

    That is an operations problem, not a teaching problem — and it is exactly the kind of problem AI automation solves. This playbook covers what AI automation actually looks like inside a Singapore tuition centre, what it costs, which government grants can fund it, and how to sequence a rollout without disrupting a live term.

    “The centres that grow are rarely the ones with the best marketing budget. They are the ones that reply to a trial-class enquiry in two minutes instead of two days.”

    What is AI automation for a tuition centre?

    AI automation for a tuition centre is a set of connected systems that handle the repetitive administrative work around teaching — enquiry response, trial booking, enrolment paperwork, fee reminders, attendance tracking, progress updates to parents, and class-change notifications — without a staff member manually doing each step.

    It is worth separating two things that often get lumped together:

    • Workflow automation follows fixed rules. A parent submits an enrolment form, so a welcome pack sends, a WhatsApp group invite goes out, and a record is created in your system. Same input, same output, every time.
    • Agentic AI handles the messy middle. A parent messages at 9pm asking “do you have P5 Math on Saturdays and how much is it?” — an AI agent reads the question, checks live class availability, answers with the correct schedule and fee, and offers a trial slot. No human, no template that only half-fits.

    Most centres need both. The rules-based layer runs the predictable admin; the AI layer absorbs the unpredictable parent communication that currently eats your front-desk staff’s day.

    The five problems worth automating first

    1. Trial-class enquiries going cold

    Parents shop around. When a parent messages three centres on a Sunday evening, the one that replies first with a concrete trial slot usually wins — not because it is better, but because it is easier. Enquiries that sit unanswered until Monday morning convert at a fraction of the rate of enquiries answered within the hour. If your enquiries arrive across WhatsApp, Instagram DMs, a website form, and the occasional phone call, and nobody owns all four channels, this is almost certainly where you are losing the most revenue.

    2. Monthly fee collection

    Chasing fees is the task every centre owner dislikes and every admin postpones. It is also perfectly suited to automation: a reminder before the due date, one on the day, and a politely escalating sequence afterwards, with outstanding balances tracked automatically. Removing the social awkwardness of the chase is often worth more to owners than the time saved.

    3. Parents feeling out of the loop

    Churn in tuition is rarely about results. It is about a parent who has not heard anything specific about their child in six weeks, decides the centre is not paying attention, and moves. Automated progress summaries — attendance, topics covered, a short tutor note — change that perception at almost no marginal cost per student.

    4. Class changes announced in group chats

    Rescheduling a class through a WhatsApp group reliably produces one parent who missed it and one student who turned up to an empty room. Targeted notifications to only the affected families, with a replacement slot offered where one exists, eliminate the whole category of complaint.

    5. Enrolment paperwork done by hand

    Manual enrolment means re-typing details, chasing missing information, and inconsistent records. A digital enrolment flow that triggers the welcome pack, the correct communication group, and the billing record the moment enrolment is confirmed removes several hours of admin per intake cycle.

    Handling student data properly: PDPA and children

    This is the part most automation vendors skip, and it matters more for tuition centres than for almost any other SME sector, because your data subjects are children.

    Singapore’s Personal Data Protection Commission issued Advisory Guidelines on the PDPA for Children’s Personal Data in the Digital Environment in March 2024. The essentials for a tuition centre:

    • A “child” means anyone aged 18 or younger.
    • Below 13: consent must come from a parent or guardian, and the parent must be told what the data will be used for.
    • Aged 13 to 17: the child can give valid consent, provided your policies are written so the child can actually understand them. If you have reason to think they do not understand, you must go to the parent.
    • Consent given before 18 remains valid once the individual turns 18.

    In practice this means your enrolment form should capture parental consent explicitly, your automated messages should go to the consented contact, and any AI system that stores conversation history or progress notes needs a defined retention period and access controls. A well-built automation makes compliance easier than a shoebox of paper forms — but only if it is designed with this in mind from the start rather than bolted on afterwards.

    Separately, if your centre teaches 10 or more students on subjects taught in mainstream schools, it must be registered with MOE under the private education framework, along with its courses and teachers. Registration runs through GoBusiness Licensing using CorpPass. Automation does not change your registration obligations, but keeping clean digital records of enrolment, attendance, and teaching staff makes any MOE-related administration considerably less painful.

    What does it cost?

    Costs vary with how many systems need connecting and how much custom logic is involved. Scope is the main driver:

    Scope What it typically includes
    Single workflow One automation — e.g. fee reminders and payment tracking, or trial-booking follow-up
    Front-of-house bundle Enquiry response agent across WhatsApp/web/Instagram, trial booking, digital enrolment, welcome sequence
    Full centre operations The above plus attendance, progress reporting, class-change alerts, fee collection, and owner dashboard

    Ongoing costs are mostly third-party: WhatsApp Business API message fees, AI model usage, and hosting for the automation platform. These scale with student volume rather than being fixed.

    The number that actually matters is the comparison against the alternative — the admin hours currently spent chasing fees and answering the same enquiries, and the enrolments lost to slow trial-class follow-up. Every centre is scoped individually and quoted against that baseline, so you can see the payback before committing. Book a free call for a scoped quote.

    Which grants can fund this?

    Enrichment and tuition centres are eligible for the standard Enterprise Singapore SME schemes, provided the business is registered and operating in Singapore with at least 30% local shareholding.

    Scheme Support level Best suited to
    PSG (Productivity Solutions Grant) Up to 50%, capped at S$30,000 per application Pre-approved solutions listed on the GoBusiness Tech Depot — e.g. a learning management or centre management system
    EDG (Enterprise Development Grant) Up to 50% for SMEs, no fixed cap Custom-built automation, process redesign, and consultancy — the right route for bespoke AI systems
    EIS (Enterprise Innovation Scheme) 400% tax deduction on qualifying innovation expenditure Reducing the net cost of qualifying development work at tax time
    SFEC (SkillsFuture Enterprise Credit) Credit offsetting out-of-pocket costs Training your tutors and admin staff to run the new systems

    Important 2026 note: the EDGE transition

    Enterprise Singapore is consolidating PSG, EDG, and MRA into a single new scheme called EDGE, launching in the second half of 2026. Under EDGE, businesses apply based on the intended activity — digitalisation, productivity improvement, market expansion — rather than choosing a grant first. Support is capped at S$100,000 per year, and the scheme opens to all Singapore businesses, not only SMEs.

    PSG, EDG, and MRA remain open until EDGE launches. If your project is ready now, there is no reason to wait; if you are planning a larger build for next year, it is worth confirming which scheme it will fall under before committing. One rule does not change across either regime: apply before you sign the contract or pay the vendor. Work already committed to is not fundable.

    Not sure which automation pays back first?

    Get a free 30-minute scoping call — we’ll map your operations and tell you honestly what to prioritise.

    Book a Free Call

    How a rollout actually works

    1. Operations mapping (week 1). We sit with your front-desk and admin staff and map what actually happens today — where enquiries arrive, who answers them, how fees are tracked, how parents are updated. Most centres discover two or three undocumented workarounds at this stage.
    2. Prioritisation and grant scoping (week 1–2). We identify which one or two automations return the most, and scope the project so it maps cleanly onto PSG, EDG, or the incoming EDGE criteria. You get a written scope you can take into a grant application.
    3. Grant application (week 2–4). Submitted before any build work begins. Approval timelines vary; we design the build so it can start the moment approval lands.
    4. Build and integration (week 4–8). We connect to what you already use — your centre management system, WhatsApp Business, payment provider, Google Workspace — rather than asking you to migrate. Nothing goes live yet.
    5. Parallel run (week 8–10). The automation runs alongside your existing process on a single class or level, with staff reviewing every outgoing message. This is where the edge cases surface: the parent with two children at different levels, the sibling discount, the makeup-class rule nobody wrote down.
    6. Go live and handover (week 10–12). Full rollout, staff training, and a written runbook. You own the system and the data.
    7. Ongoing tuning. Message wording, escalation thresholds, and reporting cadence get adjusted based on how parents actually respond over the first term.

    A single-workflow project compresses this to about four weeks. A full operations build runs closer to twelve. Either way, nothing is switched on across your whole centre until it has been proven on a subset.

    Frequently asked questions

    Will parents know they are talking to a bot?

    They should. We build AI agents that identify themselves and hand off to a human the moment a conversation goes beyond scheduling, fees, or general enquiries. Pretending an AI is a person is a bad trade — it works until it doesn’t, and when it fails it costs you the parent’s trust. A clearly-labelled assistant that answers instantly at 10pm is a better parent experience than a human who replies on Monday.

    Do I need to replace my current centre management system?

    No. Most tuition centres already run something — a centre management platform, a spreadsheet, a WhatsApp Business account, a payment link. Automation connects those rather than replacing them. Replacing a working system mid-term is disruptive and rarely necessary.

    What happens to student data? Where is it stored?

    Data stays in systems you own and control. We define retention periods, restrict access by role, and log what the automation does. Given that your data subjects are minors, we treat this as a design constraint from the first workshop rather than a compliance checkbox at the end.

    Is a tuition centre too small for this?

    A single-branch centre with 80 to 150 students is a good fit for a front-of-house bundle — that is enough enquiry and fee volume for the automation to pay back. Below roughly 50 students, a single targeted workflow (usually fee reminders or trial follow-up) makes more sense than a full build.

    How long before we see a return?

    Admin time savings show up within the first month of go-live. Enrolment improvements from faster enquiry response take a full intake cycle to measure properly — plan on one term before judging the conversion impact.

    Can this handle peak periods like PSLE or O-Level enrolment season?

    That is where it earns its keep. Enquiry volume in a peak intake window can be three to four times normal, and it is exactly when your admin staff have the least capacity to respond quickly. An automated front desk does not have a peak season.

    Where to start

    If you run a tuition centre in Singapore and enquiries are going cold, fee collection is eating your week, or parents are drifting because they feel uninformed, the fastest way to know whether automation is worth it is a scoping conversation — not a demo.

    Book a free call and we will map your current operations, tell you honestly which one or two automations would return the most, and scope it so it maps onto the grant you are eligible for. If automation is not the right answer for your centre, we will say so.

    Related reading: Government grants for AI automation in Singapore · AI automation by industry · Our services

  • AI Automation for Auto Workshops in Singapore: The 2026 Playbook

    AI Automation for Auto Workshops in Singapore: The 2026 Playbook

    Singapore’s car population has been frozen at roughly 660,000 cars since a 0% vehicle growth rate took effect in 2018, yet the number of independent repair workshops keeps climbing — an estimated 2,583 establishments as of 2024, up around 5% a year. More shops are chasing a fleet that cannot grow, electric vehicles now make up over 45% of new car registrations and need far less routine servicing, and the trade is consistently one of the most complained-about consumer categories in Singapore. Every one of those pressures rewards the workshop that runs leaner and feels more trustworthy than the one next door — and that is an operations problem AI automation is built to solve.

    This playbook covers what AI automation actually looks like inside a Singapore auto workshop, what it costs, which government grants can fund it, and how to roll it out without disrupting bays that are already busy.

    “More shops are chasing a fleet that cannot grow. The workshops that win aren’t the cheapest — they’re the ones customers trust and don’t have to chase for updates.”

    What is AI automation for an auto workshop?

    AI automation for an auto workshop is a set of connected systems that handle the repetitive front-desk and back-office work around servicing a car — booking intake, appointment reminders, job status updates, quote approvals, parts reordering, and fleet-account communication — without a service advisor manually doing each step by phone or WhatsApp.

    Two layers are worth separating:

    • Workflow automation follows fixed rules. A customer messages the workshop’s WhatsApp number, so an AI agent checks the booking calendar, offers open slots, confirms the appointment, and sends a reminder the evening before. No human touches any of those steps.
    • Agentic AI handles judgment calls inside guardrails. A customer describes a noise the car is making; the assistant asks clarifying questions, flags it as a possible brake issue, and books the right service bay and duration — then hands off to a mechanic for anything outside its rules, such as price negotiation or a warranty dispute.

    Most workshops start with workflow automation on booking and reminders, then add agentic capability once the first layer is proven.

    Why this matters for Singapore workshops specifically

    Market reality What it means for a workshop
    Car population flat at ~660,000 since 2018; private cars have declined since 2022 No fleet growth to ride — growth has to come from winning a larger share of a fixed pool of cars, not a bigger pool
    ~2,583 repair establishments, rising ~5% a year More competitors chasing the same flat demand — price alone is a losing strategy since independents already undercut dealers by design
    EVs were 45.1% of new car registrations in 2025, up from 3.8% in 2021 Routine ICE servicing volume shrinks over the next decade as the fleet turns over; workshops need to capture higher-value work (ADAS calibration, EV/HV servicing, fleet contracts) to offset it
    Motorcars are consistently among Singapore’s most-complained-about consumer categories (CASE, 2024) Trust is scarce and valuable — transparent, automated communication is a real differentiator, not just a convenience feature
    Fleet and private-hire servicing is the fastest-growing customer segment B2B accounts need reliable status updates and invoicing at volume — a good fit for automation, a poor fit for manual phone calls

    Where AI automation actually helps in a workshop

    Pain point What AI automation does
    Front desk tied up answering “got slot or not” messages all day WhatsApp booking assistant shows live availability and confirms appointments 24/7, including outside opening hours
    Customers who no-show or forget appointments Automated reminders the day before and morning of — workshops using this pattern typically cut no-shows meaningfully
    Customers anxious about being overcharged or not knowing what’s happening to their car Automated job-status updates with photos of the actual worn part, itemised digital quotes sent for approval before work starts
    Parts reordering handled reactively, causing delays mid-job Inventory automation flags low stock and triggers reorders against historical usage patterns
    Fleet/private-hire accounts calling in for status on multiple vehicles A dedicated fleet portal or WhatsApp channel with automated status pushes per vehicle, batch invoicing
    Manual data entry between the booking system, the invoicing tool, and the accountant Systems talk to each other — a completed job automatically triggers an invoice and updates the customer’s digital service record

    The trust angle: automation as a differentiator, not just an efficiency play

    Only around 27 motoring businesses in Singapore hold CaseTrust-SVTA accreditation — the formal mark of standard contracts, itemised pricing, and bonded dispute resolution. That scarcity is the opening. A workshop that automatically sends a photo of the worn part before repair, an itemised digital quote before work starts, and a written service record after completion is doing, by default, most of what CaseTrust accreditation exists to prove — and doing it for every customer without extra admin headcount.

    This also touches the Personal Data Protection Act (PDPA): a workshop’s booking and CRM systems hold customers’ names, phone numbers, vehicle registration numbers, and service history. Any automation vendor should be able to state plainly where that data is stored, who can access it, and how long it is retained — the same standard Stack AI applies to every client system.

    Government grants that can fund this

    Grant Funding Fit for a workshop
    Productivity Solutions Grant (PSG) Up to 50% co-funding, S$30,000 annual cap Covers pre-approved productivity IT solutions from the GoBusiness Tech Depot — workshop management systems, customer communication tools, and accounting/inventory software all have pre-approved options
    Enterprise Development Grant (EDG) Up to 50% co-funding Best fit for a custom-built automation project (e.g. a bespoke WhatsApp booking + fleet-portal system) rather than an off-the-shelf PSG solution, since EDG funds broader business/process transformation projects

    Standard eligibility for both applies as with any Singapore SME: the business must be registered and operating in Singapore, with at least 30% local shareholding. Workshops of any size that meet these criteria can apply — there is no automotive-specific exclusion. Apply before signing any vendor contract; retroactive funding is not available.

    Not sure where to start?

    Stack AI scopes a free assessment of your workshop’s booking, communication, and back-office workflows — and tells you honestly which parts are worth automating first.

    Book a free assessment

    How a rollout typically works

    1. Free workflow assessment. We map how bookings, reminders, quotes, and parts ordering currently happen — phone, WhatsApp, paper, spreadsheets — and flag where the busiest manual bottlenecks are.
    2. Scope the first system. Most workshops start with WhatsApp booking and reminders, since it touches every customer and shows results within weeks.
    3. Build and connect. The automation is built to work with (not replace) whatever booking or accounting software the workshop already uses, wherever practical.
    4. Run alongside existing operations. New bookings route through the system while walk-ins and phone bookings continue as normal — no forced cutover on a busy floor.
    5. Expand once proven. Once booking and reminders are working, add digital quotes, fleet-account automation, or inventory triggers based on what the workshop actually needs.
    6. Handover and support. Staff are trained to monitor and adjust the system themselves, with support on call for anything more complex.

    Frequently asked questions

    Will AI automation replace my mechanics or service advisors?

    No. It replaces repetitive coordination work — answering “got slot or not,” typing the same reminder message, chasing a customer for approval on a quote — so your staff can spend more time on cars and on customers who need a real conversation, such as a warranty dispute or a price negotiation.

    My workshop still runs mostly on WhatsApp and a paper diary. Can this still work for us?

    Yes — WhatsApp is usually the starting point, not something to replace. The automation sits on top of your existing WhatsApp Business number and calendar, so customers keep messaging the same number they already have saved.

    Does this help with the shift toward EV and ADAS work?

    Indirectly. Automation doesn’t teach a technician EV or ADAS skills, but it frees up the front-desk and admin time needed to build those higher-value service lines — and a clean digital service record is exactly what a fleet or insurer wants to see before trusting a workshop with EV/HV or ADAS calibration work.

    How much does this typically cost?

    A single workflow, such as WhatsApp booking and reminders, typically runs S$3,000–S$8,000 to build. A fuller front-of-house system covering quotes, job-status updates, and parts triggers is typically S$10,000–S$25,000. These are scope-dependent estimates, not fixed quotes — a free assessment gives you an exact number for your workshop, and PSG/EDG co-funding can offset up to half of it.

    What data does an automation system store, and is it PDPA-compliant?

    Typically customer name, phone number, vehicle registration, and service history — the same data most workshops already hold in a booking diary or CRM. Stack AI builds every system with a clear data-retention and access policy in line with PDPA, and can walk through exactly where the data sits before any project starts.

  • EIS Tax Deduction for AI Automation in Singapore: The Complete 2026 Guide

    EIS Tax Deduction for AI Automation in Singapore: The Complete 2026 Guide

    Most Singapore SMEs looking at AI automation already know to ask about the Enterprise Development Grant or the Productivity Solutions Grant. Fewer know that the Enterprise Innovation Scheme (EIS) — a tax scheme run by IRAS, not a grant run by Enterprise Singapore — can now hand back up to 400% tax deduction on AI adoption costs, on top of whatever grant co-funding a project already receives. Because EIS works through your company’s tax return rather than a grant application, it stacks with EDG, PSG, and TSS rather than competing with them.

    This guide explains what EIS is, what changed for AI in Budget 2026, who qualifies, and how it fits alongside the grant schemes already covered in Stack AI’s grant guides.

    “EIS isn’t a grant you apply for before the project starts — it’s a deduction you claim after, through your tax filing. That makes it the easiest funding mechanism to miss, and the easiest one to combine with everything else.”

    What is the Enterprise Innovation Scheme?

    The Enterprise Innovation Scheme is a Singapore corporate tax scheme, administered by IRAS, that lets businesses claim enhanced tax deductions or allowances — up to 400% — on qualifying spending across research and development, intellectual property, staff training, and (from Budget 2026) AI adoption. It runs from Year of Assessment (YA) 2024 to YA 2028. Unlike EDG, PSG, or TSS, there is no pre-approval step and no vendor pre-qualification list — the deduction is simply claimed when the company files its tax return (Form C-S or Form C), based on expenditure the business already incurred.

    Since its introduction, EIS has covered five qualifying activities: R&D undertaken in Singapore; registration of intellectual property (patents, trademarks, designs, plant varieties); acquisition and licensing of IP rights; training via SkillsFuture Singapore-approved courses or courses run with ITE and other approved partners; and innovation projects carried out with a polytechnic, ITE, or other approved partner.

    What changed in Budget 2026: AI adoption is now a qualifying activity

    Budget 2026 added a sixth qualifying activity specifically for AI adoption, available for YA 2027 and YA 2028. A business can claim 400% tax deduction on up to S$50,000 of qualifying AI expenditure per YA — meaning S$50,000 spent on qualifying AI adoption can generate up to S$200,000 in tax deductions for that year. For a company with a December financial year end, YA 2027 covers the financial year ending 31 December 2026 — so AI automation spend incurred now is very likely the spend that counts.

    Detail AI adoption category (new, Budget 2026)
    Deduction rate 400% (100% base deduction/allowance + 300% enhanced, or a full 400% where the expenditure isn’t otherwise deductible under Section 14)
    Expenditure cap S$50,000 per Year of Assessment
    Years available YA 2027 and YA 2028 only
    Cash payout option Not available for this category (cash conversion applies only to the five original categories)
    Qualifying spend (expected) Costs directly tied to adopting AI — system subscriptions/licensing, implementation and integration, and related capability-building. IRAS said a detailed list would follow; businesses should confirm the current list on IRAS’s EIS page before filing, since the exact boundaries of “qualifying AI expenditure” were still being finalised as this guide was written.

    This is the category most directly relevant to a Stack AI engagement: a WhatsApp AI agent, an invoice-processing system, or a workflow-automation build all sit squarely inside “adopting AI,” as distinct from buying generic software.

    Who is eligible

    Requirement Detail
    Entity type Singapore-incorporated companies carrying on a trade or business (sole proprietorships and partnerships can access some categories via the cash payout route)
    Activity Must fall under one of the six qualifying activities and be incurred in the relevant YA
    Application None required for the tax deduction — claimed directly in the tax return with supporting records kept for IRAS’s request
    Cash payout (legacy categories only) Optional: convert up to S$100,000 of combined qualifying expenditure per YA (excluding AI adoption) into cash at a 20% conversion rate, capped at S$20,000 cash per YA, and the payout is not taxable

    How EIS fits with EDG, PSG, and TSS

    The most common mistake is treating EIS as competing with the grant schemes rather than sitting on top of them. A grant reduces the upfront cost of a project; EIS then lets the business deduct the remaining self-funded portion at up to 400% at tax time. They’re claimed at different points in the process, from different agencies, and are not mutually exclusive.

    Scheme Run by What it does When you claim
    EDG Enterprise Singapore Up to 50% co-funding on eligible project costs Applied for before the project starts
    PSG Enterprise Singapore 50% funding on pre-approved solutions (Tech Depot) Applied for before purchase
    TSS NCSS Up to 80% co-funding for social service agencies Applied for before the project starts
    ASME-UOB package ASME + UOB Access/bridging layer on top of PSG/EDG for micro and small firms Alongside the underlying grant application
    EIS IRAS Up to 400% tax deduction on qualifying spend, incl. AI adoption from YA 2027 Claimed after the fact, in the company’s tax return

    In practice: a clinic that gets 50% EDG co-funding on a S$20,000 automation project pays S$10,000 out of pocket. If that S$10,000 of AI-adoption spend qualifies under the new EIS category, the clinic can also claim up to S$40,000 in tax deductions against that S$10,000 — on top of the grant, not instead of it.

    How to claim EIS on a Stack AI project

    1. Scope the project and identify the AI-adoption component. Not every line item in an automation build necessarily counts — the AI-specific elements (the agent, the model usage, the integration work that makes the system “AI”, as distinct from generic hosting or hardware) are what IRAS’s guidance is expected to target.
    2. Keep the invoice and project documentation itemised. IRAS doesn’t pre-approve EIS claims, but it can request supporting records, so a vendor invoice that clearly separates AI-adoption costs from other costs makes a later claim far easier to support.
    3. Confirm the current qualifying-expenditure list with IRAS or your tax agent before filing. Because the detailed criteria for the AI category were still being finalised, this is the one step in the whole EIS process that genuinely needs a check against the live IRAS page or a conversation with an accountant, not assumption.
    4. Combine with any grant already received. If the project also received EDG, PSG, or TSS co-funding, only the self-funded portion is typically what’s left to claim under EIS — your tax agent applies the deduction to the actual out-of-pocket expenditure.
    5. Claim in the tax return for the relevant YA. No separate EIS application form for the deduction itself — it’s reported as part of Form C-S or Form C when the company files.

    Frequently asked questions

    Is EIS a grant or a tax scheme?

    A tax scheme, not a grant. It’s administered by IRAS and delivered as a deduction or allowance against taxable income (or a cash conversion for the five original categories), not as a co-funding cheque before a project starts.

    Does EIS require pre-approval like PSG’s Tech Depot?

    No. There’s no pre-approved vendor list and no application before spending. The claim is made when the business files its tax return for the relevant Year of Assessment.

    Can I get a cash payout for AI adoption spending under EIS?

    No. The cash conversion option (up to S$100,000 of expenditure at a 20% rate, capped at S$20,000 cash per YA) applies only to the five original qualifying activities. The new AI-adoption category is deduction-only.

    Can EIS be combined with EDG, PSG, or TSS on the same project?

    Yes — they operate independently. A grant reduces the upfront cost; EIS is then claimed against whatever the business still spent out of pocket, at tax-filing time.

    What counts as “qualifying AI expenditure”?

    IRAS indicated the detailed list would be published separately from the initial Budget 2026 announcement. Based on public commentary so far, it’s expected to cover AI system subscriptions and licensing, implementation/integration costs, and related capability-building — but businesses should check IRAS’s EIS page directly, or ask their tax agent, before relying on this for a specific claim.

    Do I need a Singapore-registered company to claim EIS?

    Yes, for the tax deduction route. The scheme is built around Singapore corporate tax filing, so it applies to Singapore-incorporated businesses carrying on a trade or business here.

    Not sure how much of your automation project would qualify?

    Stack AI scopes projects with the funding stack in mind — grant co-funding first, EIS on top of it. Get in touch and we’ll walk through what applies to your business.

    Talk to Stack AI

  • AI Grants for Micro and Small Businesses in Singapore: The ASME-UOB S$10M Support Package (2026 Guide)

    AI Grants for Micro and Small Businesses in Singapore: The ASME-UOB S$10M Support Package (2026 Guide)

    Most government grant guides in Singapore assume a certain size of business — a registered company with a few staff, a finance person who can pull financial statements, and enough cash on hand to pay a vendor upfront and wait weeks for reimbursement. For a huge slice of Singapore’s economy, none of that fits. The one-owner clinic, the two-person marketing consultancy, the family-run F&B stall that just incorporated — these businesses make up the vast majority of local enterprises, and they are exactly the ones a new S$10 million support package announced by ASME and UOB in January 2026 is built for.

    This guide explains what the package actually is, who qualifies, how it differs from PSG, EDG, and EIS, and how a micro or small business can use it to fund AI automation.

    “For many micro and small businesses, the primary barrier to AI adoption is not a lack of ambition, but the inherent risk of digital transformation.” — Ang Yuit, President, ASME

    What is the ASME-UOB S$10 million package?

    It is not a new grant scheme with its own application portal. It is a partnership, formalised through a Memorandum of Understanding signed by UOB and the Association of Small and Medium Enterprises (ASME) on 23 January 2026 at AI Festival Asia, that does two things:

    • Ringfences access to existing government grants. ASME is mobilising up to S$10 million in support by working with government agencies to make sure the smallest businesses — the ones that often assume they are “too small” for a grant — can actually get through the door and qualify.
    • Adds a cash-flow bridge. Under the MOU, UOB provides financing so an eligible business can start a project and pay its vendor before the grant reimbursement lands, rather than having to fund the whole project out of pocket first and wait.

    The scoping and matching work runs through SME@AITE, a joint AI Centre of Excellence set up by ASME and the Institute of Technical Education, which helps SMEs define a project and connects them with implementation partners and, where relevant, ITE students working on real-world AI prototypes through a parallel S$1 million ASME–Lenovo initiative called AI Foundry.

    Who is it for?

    The package is aimed squarely at micro and small enterprises — reported at the time of announcement as businesses with 30 or fewer employees. Singapore’s micro and small enterprises make up roughly 94% of all business entities in the country and employ close to half the workforce, so this is a large and previously under-served segment.

    In practice, this is squarely Stack AI’s smallest-end client profile: a solo GP clinic, a single-location tuition centre, a two- or three-person insurance or financial advisory practice, an independent gym, a small auto workshop, a family-run F&B outlet that has only just incorporated. These businesses are technically eligible for PSG and EDG today, but many self-select out — they assume a grant application needs a finance department, or they cannot absorb paying a vendor upfront while waiting 8–12 weeks for reimbursement. This package is aimed directly at removing those two blockers.

    How it compares to PSG, EDG, and EIS

    The ASME-UOB package is best understood as sitting alongside the existing schemes Stack AI projects already qualify under — not replacing them.

    Scheme What it is Support level Where the ASME-UOB package helps
    PSG Grant for pre-approved off-the-shelf solutions Up to 50%, capped at S$30,000/year Ringfenced access + cash-flow bridge via UOB financing
    EDG Grant for custom projects (Process Redesign / Automation) Up to 50%, no fixed cap ASME scoping support via SME@AITE; UOB bridging loan pre-disbursement
    EIS Tax deduction / cash payout on innovation spend 400% deduction (or 20% cash payout if loss-making) Not directly affected — claimed at tax filing regardless
    ASME-UOB package Access + financing layer, not a standalone grant Up to S$10M in existing grants mobilised, plus UOB loans The mechanism that makes PSG/EDG practically reachable for the smallest firms

    The one thing that does not change under any of these routes: you must apply, and get approval, before signing a contract or paying a vendor. A project that has already started is not fundable retroactively — this rule holds across PSG, EDG, and the ASME-UOB pathway alike.

    What this means in dollar terms for a small AI automation project

    Take a solo-practitioner clinic or a two-person advisory firm considering a S$12,000 front-of-house automation build (WhatsApp enquiry handling, appointment booking, follow-up sequences):

    Step Without the package With the ASME-UOB package
    Upfront cash needed Full S$12,000, paid before grant reimbursement arrives Bridged by UOB financing while the EDG/PSG claim is processed
    Grant support Up to 50% via EDG if the business can navigate the application alone Same 50%, plus ASME/SME@AITE help with scoping and the application itself
    Net out-of-pocket after grant ~S$6,000, all paid upfront ~S$6,000, spread across the financing period instead of paid in one lump sum

    The grant percentage does not change — what changes is whether a business with thin cash reserves can actually get to day one of the project without the upfront cost stalling it.

    Not sure if you qualify?

    We’ll tell you honestly whether your business fits PSG, EDG, or this new pathway — before you spend a cent.

    Book a Free Call

    How to access it: step by step

    1. Check ASME membership. The ringfenced grant access and SME@AITE support run through ASME — confirm membership status or sign up via asme.org.sg first.
    2. Scope the project. Talk to us (or another implementation partner) about what the automation should actually cover. A clear, right-sized scope is what SME@AITE and any grant assessor will want to see — not a vague “we want AI.”
    3. Get matched via SME@AITE. ASME’s AI Centre of Excellence helps confirm which existing grant (typically PSG for pre-approved solutions, EDG for custom builds) the project should apply under, and flags where UOB financing can bridge the gap.
    4. Apply for the underlying grant before starting. The application still runs through the standard Business Grants Portal (EDG/PSG) — the ASME-UOB layer smooths access and cash flow, it does not replace the application step.
    5. Arrange UOB financing if needed. Once the grant application is in motion, eligible businesses can discuss bridging finance with UOB to cover costs before reimbursement lands.
    6. Build, go live, claim. The project proceeds as any grant-funded automation build would — and the reimbursement, once received, pays down the bridging finance.

    Frequently asked questions

    Is this a new grant I apply for directly?

    No. There is no separate “ASME-UOB grant” application form. It is an access and financing layer on top of existing schemes (mainly PSG and EDG) — you still apply through the normal Business Grants Portal process, with ASME’s SME@AITE helping with scoping and UOB providing optional bridging finance.

    Do I need to bank with UOB already?

    The financing component is delivered by UOB, so accessing it will involve a UOB business banking relationship. The grant-access and scoping support through ASME/SME@AITE is separate from that and does not require an existing UOB relationship.

    What counts as a “micro or small enterprise” here?

    Reported coverage of the announcement defines the target group as businesses with 30 or fewer employees. ASME has not published a separate formal legal definition specific to this package as of this guide — if your business is close to that threshold, confirm directly with ASME before assuming eligibility either way.

    Does this replace EDG or PSG eligibility criteria?

    No. The underlying grants’ own rules still apply — Singapore registration, local shareholding thresholds, and the “apply before you start” rule. The package addresses practical barriers (knowing how to apply, affording the wait for reimbursement), not the formal eligibility criteria themselves.

    Is this only for AI projects, or any digitalisation spend?

    The announcement frames it around AI adoption specifically — AI-enabled tools, and hiring or training staff to work with them — rather than digitalisation in general. An AI automation project, like the workflow and agentic systems Stack AI builds, sits squarely inside its intended scope.

    When did this start, and is it still running?

    The MOU was signed 23 January 2026. As of this guide, officials have indicated further public initiatives for small and micro firms would follow in the coming weeks and months — worth checking with ASME for the latest status before assuming details are unchanged.

    Where to start

    If your business is small enough that you have assumed government grants are not really built for you, this is worth a second look. Book a free scoping call and we will map what an automation project would look like for your business, tell you honestly which funding route fits, and help you sequence it so the grant application happens before any work starts — the one rule that never changes.

    Related reading: Government grants for AI automation in Singapore · AI automation by industry · Our services

  • What is the Enterprise Development Grant (EDG) — and how do you actually qualify?

    What is the Enterprise Development Grant (EDG) — and how do you actually qualify?

    If you run an SME in Singapore and you’ve started looking into government grants for technology or automation projects, you’ve almost certainly come across the Enterprise Development Grant (EDG). It’s the government’s flagship scheme for helping local businesses invest in upgrading their capabilities — and for automation and AI projects specifically, it’s often the most relevant grant available.

    But most business owners either don’t know if they qualify, aren’t sure what costs it actually covers, or find the application process opaque enough that they give up before starting. This guide changes that.

    What is EDG?

    The Enterprise Development Grant is administered by Enterprise Singapore (EnterpriseSG) — the government agency responsible for helping Singapore SMEs grow and upgrade. EDG is designed to support companies that want to invest in building stronger core capabilities, improve their operations, or expand into new markets.

    It’s a co-funding scheme: EnterpriseSG pays a portion of your qualifying project costs, and you cover the rest. The grant doesn’t go directly to you — it offsets the fees you pay to an approved vendor or consultant who delivers the project.

    Key figure Qualifying Singapore SMEs can receive up to 50% co-funding on eligible project costs under EDG. This means for every $10,000 you spend on a qualifying automation project, the government co-funds $5,000.

    EDG was introduced in 2018 by merging several older grant schemes into a single, more flexible programme. It has since become the primary tool Singapore companies use to fund digitalisation, automation, AI adoption, process improvement, and market expansion initiatives.

    The three pillars of EDG

    EDG is organised around three broad areas of business development. Every project you apply for must fall under one of these pillars:

    Pillar What it covers
    Core Capabilities Building foundational business strengths — strategy, financial management, human capital development, service excellence, and brand development.
    Innovation & Productivity Upgrading processes, adopting technology, automating workflows, and implementing AI and digital solutions. This is the pillar most automation projects fall under.
    Market Access Expanding into overseas markets — business development, market entry activities, and international partnerships.

    For the vast majority of AI automation and workflow projects, you’ll be applying under Innovation & Productivity. This pillar covers technology adoption, process redesign, and productivity improvements — which is exactly what a well-scoped automation project delivers.

    Who qualifies for EDG?

    This is where most business owners get tripped up, because the criteria are specific and non-negotiable. You need to meet all of the following:

    EDG Eligibility Criteria

    • Registered and operating in Singapore. Your business must be incorporated and physically operating in Singapore — not just registered here as a holding entity.
    • At least 30% local shareholding. Singapore Citizens or Permanent Residents must hold at least 30% of the company’s shares. Wholly foreign-owned subsidiaries do not qualify.
    • Financially viable. EnterpriseSG will assess your company’s financial health. A business that is loss-making for several consecutive years, or has significant liabilities, may be declined.
    • Project must be commercially meaningful. The work must have a clear impact on your business — not a proof-of-concept that you won’t actually deploy.
    • Work must be done by a third-party vendor. You cannot claim EDG for internal staff doing the project themselves. The work must be performed by an external company (like Stack AI).
    • Project must not have started before approval. This is critical. If you begin work before your EDG application is formally approved, those costs become ineligible. Always apply first.

    There is no official “SME only” rule written into EDG’s base criteria, but the 50% co-funding rate applies to SMEs — defined as companies with annual sales turnover not exceeding S$100 million, or fewer than 200 employees. Larger enterprises may still apply but typically receive a lower support percentage.

    What costs does EDG actually cover?

    Understanding what’s fundable is just as important as knowing whether you qualify. EDG covers three categories of qualifying costs:

    1. Professional fees

    This is the main cost category for automation and technology projects. It covers the fees charged by your implementation partner — in this case, the fees Stack AI charges for scoping, building, testing, and deploying your automation system.

    2. Software and equipment costs

    Costs for new software licences, subscriptions, or equipment directly required for the project may be included. For example, if your automation project requires a new Make.com subscription or a specific AI platform licence, that may be claimable.

    3. Internal manpower costs

    If your own staff are significantly involved in the project — attending workshops, doing data preparation, project management — a portion of their time can sometimes be included as a qualifying cost. This is the most complex category and requires good documentation.

    What this means in practice For a typical Stack AI automation project costing S$15,000, an eligible SME could receive up to S$7,500 back via EDG co-funding — reducing the net cost to S$7,500. For larger enterprise-grade projects, the absolute savings are proportionally higher.

    What EDG does NOT cover

    Equally important is what you cannot claim. Common exclusions include:

    Not Covered by EDG

    • Not eligible: Recurring operational costs. Ongoing SaaS subscriptions, monthly hosting fees, or staff salaries after the project is complete are not claimable. EDG funds project-based work, not running costs.
    • Not eligible: Work done before approval. Any invoices dated before your Letter of Offer from EnterpriseSG are automatically excluded — even if your approval comes through quickly.
    • Not eligible: Hardware not specific to the project. General IT equipment like laptops or office computers don’t qualify unless they are specifically and exclusively required for the funded project.
    • Not eligible: Training-only programmes. Sending staff to generic technology training without a tied implementation project is not fundable under EDG’s Innovation & Productivity pillar (though it may qualify under other schemes like SkillsFuture).
    • Not eligible: Projects with insufficient scope. Very small projects — or projects that EnterpriseSG deems too narrow to meaningfully improve your business — may be declined. A minimum project investment is typically expected.

    How much can you actually get?

    The current co-funding levels for EDG are:

    • SMEs (up to S$100M turnover or under 200 employees): Up to 50% of qualifying costs
    • Non-SME companies: Up to 30% of qualifying costs

    Note that EnterpriseSG assesses each project independently and may approve a lower percentage based on factors like the strength of your application, how clearly you’ve articulated the business impact, and your company’s financial position. The 50% figure is the ceiling, not a guarantee.

    There is also a maximum qualifying cost per project — EnterpriseSG sets internal caps and may not fund the full cost of very large projects at the maximum rate. For most SME automation projects in the S$10,000–S$80,000 range, the 50% rate is achievable with a well-prepared application.

    Why automation projects are well-suited for EDG

    Automation and AI implementation projects score well with EnterpriseSG assessors because they align perfectly with what EDG’s Innovation & Productivity pillar is designed to fund. Specifically:

    • They have clear, measurable outcomes. A good automation project comes with defined metrics: hours saved per week, reduction in error rate, faster turnaround times. These are exactly what EnterpriseSG assessors want to see.
    • They upgrade capability, not just spend. EDG is about building business capability, not purchasing off-the-shelf software. A custom automation system built for your business demonstrates genuine capability uplift.
    • They’re commercially deployed. EnterpriseSG wants to fund projects that go live and generate real impact — not R&D experiments. Automation projects that go straight into production score well.
    • Singapore’s national AI strategy aligns. The government has made AI and automation a national priority. EDG assessors are primed to view well-scoped AI projects favourably.

    The key is how the project is scoped and presented. A poorly framed application — even for a genuinely impactful project — can result in a lower approval rate or a rejection. The way you describe the business problem, the expected productivity gains, and the implementation approach all affect the outcome.

    Not sure if your project qualifies?

    Not sure if your project qualifies? We scope automation projects in grant-eligible terms from day one — and we’ll tell you honestly if EDG is the right fit before you invest any time in the application.

    Book a Free Call

    The application process, step by step

    The EDG application is submitted through the Business Grants Portal (BGP) at businessgrants.gov.sg. Here’s what the process looks like in practice:

    1. Scope your project with your vendor — Before applying, work with your implementation partner to define the project scope, timeline, and cost. You’ll need a formal quote and a clear statement of the business problem being solved. This is where most of the work happens — a vague scope produces a weak application.
    2. Submit your application on the BGP — Log into the Business Grants Portal using your CorpPass credentials. Select “Enterprise Development Grant” and complete the application form. You’ll describe your business, the project, expected outcomes, and attach your vendor’s proposal and quote. Allow 1–2 hours to complete carefully.
    3. Wait for assessment (typically 4–8 weeks) — EnterpriseSG reviews your application. They may come back with clarifying questions — respond promptly and clearly. Assessment timelines vary based on application volume and project complexity. Do not start work during this period.
    4. Receive your Letter of Offer (LOO) — If approved, you’ll receive a Letter of Offer specifying the approved funding amount and conditions. Read this carefully — it defines exactly what’s covered and the timeline you must complete the project within (typically 12–18 months).
    5. Accept the offer and start work — Once you formally accept the LOO via the BGP, you can begin the project. All work from this point forward is eligible for co-funding. Keep all invoices, timesheets, and project documentation — you’ll need them for the claims process.
    6. Complete the project and submit your claim — After the project is delivered, you submit a claims package via the BGP. This includes invoices, proof of payment, and evidence that the project was completed as scoped (reports, screenshots, outcome documentation). EnterpriseSG reviews and disburses the co-funding — typically within 2–3 months of a complete claim submission.

    Common mistakes that get applications rejected

    After working with multiple Singapore SMEs on grant-funded projects, here’s what we see go wrong most often:

    • Starting work before approval. This is the single most common and most avoidable mistake. If any invoice is dated before your Letter of Offer, that cost is ineligible — full stop.
    • Vague project scope. “We want to automate our business” is not a project scope. You need to specify: which workflows, which systems, what the current process looks like, what the automated process will look like, and what measurable improvement you expect.
    • Not quantifying the business impact. EnterpriseSG wants to fund projects that genuinely improve Singapore’s economic competitiveness. Applications that don’t include concrete projections — hours saved, cost reduction, revenue impact — are much weaker than those that do.
    • Using a vendor who can’t provide proper documentation. Your vendor must provide a formal proposal, detailed quote, and be able to produce project completion evidence. Informal arrangements or freelancers without proper invoicing will create problems at the claims stage.
    • Applying for costs that won’t be approved. Claiming ineligible items (like existing staff salaries or software you were already paying for) doesn’t just get those items rejected — it can raise questions about the whole application.

    How Stack AI works with you on EDG

    We’ve scoped and delivered automation projects with EDG co-funding for clients across F&B, professional services, healthcare, and retail. Here’s how we approach it:

    We scope projects in grant-eligible terms from day one. This means defining clear business problems, mapping current workflows, specifying deliverables, and projecting measurable outcomes — all in language that holds up under EnterpriseSG scrutiny.

    We provide everything you need for the application. Our proposals include the technical scope, timeline, cost breakdown, and expected business impact in a format that works directly as your BGP submission attachment. You don’t need to translate our work into grant language — we do that as part of how we scope every project.

    We only take on projects we believe will be approved. We’re not a grant consultancy and we don’t guarantee approval — EnterpriseSG makes that decision. But we won’t encourage you to apply if we think the project isn’t a strong fit for EDG. A rejected application wastes everyone’s time.

    We stay involved through the claims process. After project delivery, we help you compile the claims documentation — usage evidence, outcome reports, and completion certification — so the final disbursement goes smoothly.

    Not sure if your project qualifies?

    Book a free 30-minute call. We’ll confirm which grants you qualify for, scope your project in grant-eligible terms, and give you a realistic picture of what you can claim — before you commit to anything.

    Book a Free Call

    Frequently asked questions

    Can I apply for EDG and another grant at the same time?

    Yes, in some cases. EDG cannot be stacked with other EnterpriseSG grants for the same cost items — you can’t double-claim the same invoice. However, if you have separate projects, you can have multiple concurrent applications. EDG can also sometimes complement schemes like the Productivity Solutions Grant (PSG) if the projects are genuinely distinct. Speak to an EnterpriseSG advisor or your vendor before stacking.

    Is there a minimum or maximum grant amount?

    There’s no hard published minimum, but EnterpriseSG expects projects of meaningful scale — very small projects (under ~S$5,000 in qualifying costs) are unlikely to be approved as they don’t demonstrate genuine capability uplift. For the maximum, there are internal caps per project and per company within a certain period, but these are not publicly specified and vary by industry and project type. For typical SME automation projects, the practical range is S$5,000–S$40,000 in co-funding per project.

    How long does the whole process take from application to receiving funds?

    Plan for 6–12 months from initial application to final disbursement. Assessment typically takes 4–8 weeks. The project itself then needs to be delivered (usually 2–4 months for most automation projects). After project completion, you submit your claim, and disbursement typically follows within 2–3 months. You pay your vendor in full first — the grant reimbursement comes after the project is done and the claim is processed.

    What if I’ve already started my project — can I still apply?

    Unfortunately, no — at least not for costs already incurred. If you’ve already paid invoices or started work, those costs are ineligible. However, if the project is ongoing, you may still be able to apply for remaining costs that haven’t been incurred yet, as long as those future phases are clearly delineated in your application. Speak to EnterpriseSG directly or contact us — we can help you assess what’s still salvageable.

    Does my vendor need to be pre-approved by EnterpriseSG?

    Not necessarily. Unlike PSG (which requires vendors to be on an approved vendor list), EDG does not require your implementation partner to be pre-approved. However, EnterpriseSG will assess the vendor’s credibility, experience, and whether their proposal is credible. A vendor with a track record of delivering similar projects and proper documentation will always strengthen your application.

    What happens if my project costs more than the approved amount?

    You cover any costs above the approved amount. If your project ends up costing more than originally scoped, EnterpriseSG only funds up to the approved quantum. If there are legitimate scope increases, you can sometimes submit a variation request — but this must be approved before those additional costs are incurred. This is another reason to work with a vendor who scopes projects carefully upfront.

    The Enterprise Development Grant is one of the most accessible and generous co-funding schemes available to Singapore SMEs — but only if you approach it correctly. The business owners who benefit most are those who plan the project before applying, work with a vendor who understands grant requirements, and take the application seriously as a business document rather than a form to fill in.

    If you’re considering an automation project and want to understand whether EDG makes sense for your situation, the best next step is a conversation — not more reading.