Singapore SME AI Adoption Hit 23.4%: What IMDA's New Report Means for Your Books

IMDA's latest Digital Economy Report shows nearly one in four Singapore SMEs used AI in 2025. Larger firms are at seven in ten. Here's why the gap is mostly about data, what AI can realistically do for a small finance team, and a 10-point AI-Ready Books Scorecard.

TL;DR

• SME AI adoption rose to 23.4% in 2025, from 14.5% in 2024, according to IMDA's Singapore Digital Economy Report 2026, released on 5 October. Among larger firms it's 70.4%.

• About three in four SMEs still don't use AI. Among firms that do, 87.6% say it improved productivity and processes.

• Nearly nine in ten AI users rely on off-the-shelf tools. For most SMEs, AI will arrive inside software they already use, not as a custom project.

• AI can only answer what your data can answer. If sales, stock and payments live in separate spreadsheets, the AI sees fragments.

• GST InvoiceNow is quietly laying the groundwork. It moves invoices into structured data for every GST-registered business between 2028 and 2031, which is exactly what AI needs.

• Going to SMEs Go Digital Day on 13 October? Score yourself on the AI-Ready Books Scorecard first, and take our 7 questions for any AI vendor.

Accurate as of 6 October 2026, based on IMDA's Singapore Digital Economy Report 2026 (released 5 October 2026) as reported by The Straits Times and other outlets, and IRAS's GST InvoiceNow announcements. Survey figures describe 2025 unless stated otherwise.

"Can you get me that by Thursday?"

It's Monday. The boss asks two simple questions. Which customers have owed us money for more than 60 days? And how many units of our best seller are actually left?

The finance person nods. "Give me till Thursday."

Nobody is being lazy. The receivables are in the accounting software. The stock is in a spreadsheet the warehouse updates "most days". The customer who paid by PayNow last week hasn't been matched yet, because the payment reference just says "inv". So two simple questions mean exporting, a VLOOKUP, a phone call to the warehouse, and trusting someone's memory.

Then somebody in the meeting says, "We should get AI for this."

The instinct is right. The starting point is wrong. And IMDA's new numbers, published on 5 October, explain why so many Singapore SMEs are stuck right there.

What IMDA's Digital Economy Report 2026 found

IMDA's Singapore Digital Economy Report 2026, released on 5 October 2026, found that 23.4% of Singapore SMEs used AI in 2025, up from 14.5% in 2024. Adoption among larger firms rose from 62.5% to 70.4%. Across all enterprises, it rose from 14.7% to 23.5%.

Source: IMDA Singapore Digital Economy Report 2026, as reported by The Straits Times, OpenGov Asia and Tech Edition, 5 to 6 October 2026.

A few more findings from the report are worth knowing:

• Benefits: among AI adopters, 87.6% cited productivity and process improvement, and 44% cited lower costs or better use of resources.

• Off-the-shelf wins: close to nine in ten AI-using firms use commercial, ready-made AI tools. 28.3% customise them.

• Going deeper: about two in five AI users are expanding AI across the business or have fully integrated it.

• Big differences by industry: information and communications led at 46.8%, then professional services (34.5%) and finance and insurance (27.6%). Construction was lowest at 7.6%.

• Skills gap: about 68% of workers say they need to build AI skills, but only 37% had attended AI training in the past year.

Two things stand out. First, the gap: a larger firm is roughly three times as likely to use AI as an SME. Second, the direction: SME adoption jumped about nine percentage points in a single year, and the share of non-users planning to start has doubled since 2023. For most owners, the question has quietly moved from "should we?" to "where do we start?"

Why most SMEs still aren't using AI

According to a Southeast Asia Connect briefing on IMDA's data, the leading barrier among firms that hadn't adopted AI was that they didn't see its relevance at their size.

That's a fair reaction. A lot of what gets sold as "AI for business" is chatbots, content generators and dashboards built for companies with a data team. A 12-person distributor doesn't need a chatbot. It needs to know who owes it money.

There's a second reason that surveys rarely capture: AI needs something to read. Larger firms tend to run integrated systems where sales, purchasing, stock and the ledger share one database. Many SMEs run on a patchwork: an accounting package for the books, Excel for stock, WhatsApp for orders, and an inbox full of PDF supplier bills. An AI sitting on top of that patchwork only ever sees fragments.

Anyone who has helped a small company move off spreadsheets will recognise the pattern. The same customer saved three different ways. Stock counts that don't match the ledger. A year of supplier bills that exist only as email attachments. None of these is an AI problem, but every one of them stops an AI from giving a straight answer.

In our EDGE Grant guide we covered MTI's August study, which suggested the bigger gains from AI come after the groundwork is in place. IMDA's new numbers add a practical detail: almost all adopters use off-the-shelf tools. Off-the-shelf tools don't come with a data team to clean up first. Your data has to be ready before they arrive.

Where AI actually helps a small finance team

For most SMEs, the realistic first step is AI built into the accounting or ERP system they already use, answering plain-English questions about their own sales, purchases, stock and cash, with a person checking the result. A standalone AI tool rarely comes first.

Look at the third column. In every row, that's the hard part. The AI is the easy part.

What "AI-ready books" actually means

You don't need a data scientist. You need four habits:

1. One source of truth. Sales, purchases, inventory and the ledger share one database, or are properly integrated. They aren't exported and re-imported every Friday.

2. Captured once. Each transaction is entered once, where it starts, and flows through. Every re-type creates a second version of the truth.

3. Structured, not scanned. An invoice stored as data (customer, item, quantity, GST code) can be read and counted. A PDF sitting in an inbox mostly can't.

4. Current. The bank is reconciled at least monthly and receivables update as payments land. An AI working on last quarter's numbers answers last quarter's questions.

GST InvoiceNow is quietly doing half the work

Here's the part most AI articles miss. If you're GST-registered in Singapore, the government is already pushing your invoices towards exactly the kind of data AI needs.

Under the GST InvoiceNow Requirement, GST-registered businesses must send invoice data to IRAS through the InvoiceNow network, which is built on the international Peppol standard. Invoices travel as structured data rather than PDFs. Existing GST-registered businesses join in phases from 1 April 2028 to 1 April 2031. (What is InvoiceNow?)

Already in: newly incorporated companies registering voluntarily for GST since 1 Nov 2025, and all new voluntary GST registrants since 1 Apr 2026. Source: IRAS, Committee of Supply 2026 factsheet (26 Feb 2026).

IRAS said in February that more than 63,000 businesses were already on the InvoiceNow network, with about 90,000 more expected to join. SMEs can use InvoiceNow-Ready Solutions free of charge until March 2031, and a new grant of up to S$1,000 was announced to help SMEs with onboarding costs.

Now look at what GST InvoiceNow asks you to get right: clean customer and supplier records, correct GST codes, invoices issued from a system instead of a spreadsheet template, and supplier bills recorded as data rather than PDFs (see our guide to the purchase side). That's the same list as "AI-ready books". Do the work once and you get both: compliance, and data an AI can actually use.

The AI-Ready Books Scorecard

Run this with your finance team or your accountant. Score one point for each "yes". You're welcome to share it.

1. One system. Sales, purchases, stock and accounts are in one system, or properly integrated, not in separate files.

2. Keyed once. Each invoice is entered once. Nobody re-types it into a second system.

3. Top customers, fast. You could name your top 10 customers by revenue this year in under five minutes.

4. Ageing you trust. Your receivables ageing is accurate to within a week.

5. Bank reconciled. Every bank account is reconciled at least monthly.

6. One record each. Every customer, supplier and item has a single record. No "ABC Pte Ltd" and "A.B.C. Pte. Ltd." side by side.

7. System-applied GST. GST codes are applied by the system, not typed in by hand.

8. Bills as data. Supplier bills are recorded as data, not left as PDFs in an inbox.

9. Stock without walking. Someone can tell you the stock on hand for any item without going to the warehouse.

10. InvoiceNow-ready. Your business has a Peppol ID and an InvoiceNow-Ready Solution, or a plan and a date to get one.

Points scoring:

• 8 to 10: you're ready. AI on top of your data can start answering real questions this month.

• 5 to 7: foundation in progress. Fix items 1, 2 and 6 first. They cause most of the wrong answers.

• 0 to 4: hold off on buying an AI tool. Fix the books first, and the AI decision becomes simple.

Going to SMEs Go Digital Day on 13 October? 7 questions for any AI vendor

IMDA's SMEs Go Digital Day 2026 is on Tuesday, 13 October, from 9am to 6pm at Suntec Singapore Convention & Exhibition Centre, Halls 401 to 403. Admission is free with registration, and the organisers list more than 250 digital solutions on the floor. If AI is on your list, take these questions with you:

1. What data does the AI read?
   Does it work on data already in your system, or will I be exporting and uploading files?

2. Can it show me the records behind an answer?
   If it says a customer owes S$12,400, can I click through to the invoices?

3. What happens when it's wrong?
   Who checks the output, and how do I correct it?

4. Where is my data stored, and is it used to train models for anyone else?

5. Is the solution InvoiceNow-Ready?
   If you're GST-registered, you'll need that between 2028 and 2031 anyway.

6. What does setup demand from my team?
   How many days of clean-up and migration before it's useful?

7. What's the grant position?
   PSG closed on 29 September. Ask how the solution fits under the EDGE Grant, and get it in writing before you pay anything.

A quick note on grants

IMDA reported that SMEs which adopted AI-enabled, pre-approved digital solutions under the Productivity Solutions Grant saw average cost savings of 49.5% in 2025. PSG has since closed. From 30 September 2026, support for digital solutions moved to the EDGE Grant, where single-function digital solutions, integrated enterprise systems and selected automation share a S$30,000 limit within the yearly cap. Our EDGE Grant guide walks through the rules and has a Grant-Ready Checklist.

Frequently asked questions

• What percentage of Singapore SMEs use AI?
  23.4% of Singapore SMEs used AI in 2025, up from 14.5% in 2024, according to IMDA's Singapore Digital Economy Report 2026, released on 5 October 2026.

• How does SME AI adoption compare with larger companies?

  Larger firms are far ahead. 70.4% of them used AI in 2025, up from 62.5% in 2024. That makes a larger firm roughly three times as likely to use AI as an SME.

• Why are Singapore SMEs slower to adopt AI?

  Among firms that haven't adopted AI, the leading barrier in IMDA's data was not seeing its relevance at their size. In practice, scattered data is another obstacle: AI can only answer questions that the business's own records can support.

• What is the easiest way for an SME to start using AI?

  Start with one question you ask often, such as which customers are overdue, and use AI built into the accounting or ERP system you already run. Close to nine in ten AI-using firms in Singapore use off-the-shelf tools rather than custom builds.

• Does GST InvoiceNow have anything to do with AI?

  Not directly. GST InvoiceNow is a tax requirement for GST-registered businesses, phased in from 2028 to 2031 for existing businesses. But it moves invoices into structured data, which is what AI tools need to answer questions reliably.

• When is SMEs Go Digital Day 2026?

  Tuesday, 13 October 2026, from 9am to 6pm at Suntec Singapore Convention & Exhibition Centre, Halls 401 to 403. It is organised by IMDA and admission is free with registration.

• Can I get a grant for AI software in Singapore?
  PSG closed on 29 September 2026. From 30 September, digital solutions are supported under the EDGE Grant, subject to its activity list and rules. Separately, SMEs can use InvoiceNow-Ready Solutions free of charge until March 2031.

Start with books an AI can read

Most "Thursday problems" aren't solved by adding another tool. They're solved by having one place where the numbers live.

AveriWorks is an integrated cloud ERP for Singapore SMEs. Accounting, sales, purchases, inventory and GST sit on the same data. It's on IMDA's list of free-of-charge InvoiceNow packages for GST-registered businesses, and it's built by Zetta Solution, an IMDA-accredited InvoiceNow-Ready Solution Provider and Access Point Provider.

• One database, so the answers add up. Invoices, stock, payments and the ledger are connected, so your receivables report and your bank agree.

• Built for what comes next. Because your sales, stock and ledger sit in one database, AveriWorks is ready for AI that can answer questions about your own business. We're working on it now.

• InvoiceNow and GST InvoiceNow built in. The compliance work and the AI groundwork happen in the same place.

• Start free. The Starter plan is free and includes accounting, unlimited InvoiceNow and GST InvoiceNow. (pricing).

Score yourself on the scorecard first. Then let the software do the Thursday work on Monday.

Start free with AveriWorks

For bigger projects, email sales@zetta-solution.com

Sources

Accurate as of 6 October 2026. Survey figures, grants, schemes and event details can change, so please check the official pages before acting on them. This article is general information and not tax, legal or financial advice.

1. IMDA, Singapore Digital Economy Report 2026 (released 5 Oct 2026): AI adoption, benefits, sector rates, worker AI use, digital economy size.

2. The Straits Times, via BritCham Singapore, AI adoption among businesses in Singapore rises; digital economy accounts for larger share of GDP (6 Oct 2026).

3. OpenGov Asia, Singapore's Digital Economy Reaches 19.3% as AI Adoption Deepens (5 Oct 2026).

4. Tech Edition, Singapore's digital economy reaches S$144.1 billion as enterprise AI adoption rises (Oct 2026).

5. Southeast Asia Connect, Briefing: What Singapore's SME AI adoption gap means for operators (5 Oct 2026): leading barrier among non-adopters.

6. IRAS, Committee of Supply 2026: Extension of GST InvoiceNow Requirement to All GST-registered Businesses by April 2031 (26 Feb 2026): timeline, network size, free solutions, S$1,000 grant.

7. IMDA, SMEs Go Digital Day 2026 registration page: date, venue, 250+ solutions.

8. Enterprise Singapore, EDGE Grant: closure of PSG, S$30,000 digital solutions limit.

9. AveriWorks, Pricing: plan contents and prices (checked 6 Oct 2026).