Guide library
Getting started (3)
What is AI readiness?What does an AI readiness assessment involve?AI readiness glossaryBuilding the foundations (5)
Process mapping before automationWhen has an SME outgrown spreadsheets?How to choose business software that fitsHow to prepare your data for AIDo SMEs need a data warehouse?By market (5)
AI readiness for Singapore SMEsAI readiness for Malaysian SMEsAI readiness for Indonesian SMEse-Invoicing in Malaysia: a guide for SMEsAplikasi stok barang untuk UMKM (Bahasa Indonesia)The framework
SEL Foundation IndexAll guides →Resources · Indonesia guide
AI readiness for Indonesian SMEs: a practical guide
Indonesia’s micro, small and medium businesses, known as UMKM, employ most of the country’s workforce. Many are now trying AI tools for marketing and customer messages. Fewer use AI in the operations that drive their margins. This guide sets out where things stand, why scattered stock and order records hold back AI, six practical steps towards using AI, and where the rules and support fit in.
The short answer
What readiness means for an Indonesian SME.
For an Indonesian SME, AI readiness means being prepared to use AI in real business operations, not only for marketing and messaging, and to get a lasting return from it.
That requires a clear business case, a senior sponsor, staff with the right skills, and lawful handling of personal data. It also requires three operational foundations: defined processes, systems that connect, and reliable data. Our guide to what AI readiness is covers each of these areas in more detail.
In Indonesia, SMEs are known as UMKM, short for usaha mikro, kecil dan menengah (micro, small and medium enterprises). Government Regulation 7 of 2021 defines them by annual sales: up to Rp2 billion for a micro business, up to Rp15 billion for a small business, and up to Rp50 billion for a medium business. This guide uses “SMEs” and “UMKM” for the same group.
Many growing Indonesian businesses still run parts of their operations through messaging apps, spreadsheets and paper. These methods work at a small scale. In our view, they become the main obstacle to AI as the business grows, because the records they leave behind are incomplete and hard to combine.
Where things stand
UMKM are most of the economy.
Government figures show how much of Indonesia’s economy depends on UMKM. A new regulation also sets a date that every business holding personal data should plan for.
of GDP comes from UMKM
The Ministry of UMKM puts the number of UMKM at around 65 million businesses.
of employment is in UMKM
Most Indonesian workers are employed by micro, small and medium businesses.
PP 33 of 2026 takes effect
The regulation implementing the Personal Data Protection Law sets detailed duties on breach notification, impact assessments and data protection officers.
Sources: Ministry of Micro, Small and Medium Enterprises; Government Regulation 33 of 2026. The UMKM shares are the figures the Ministry continues to cite. Figures as published at the time of writing.
There is no official figure yet for how many UMKM use AI, and the surveys that exist disagree widely. In our view, the more useful question for a growing business is whether its processes, systems and data could support AI in daily operations.
Stock and orders
Why scattered stock and order records hold back AI.
Many Indonesian distributors, traders and shops take orders through messaging apps, track stock in a spreadsheet, and record payments in a separate accounting system. Each part works on its own. The problem is that no single record shows the whole transaction, from the order to the delivery to the payment.
This causes problems every day, before AI is even considered. Stock figures in the spreadsheet drift away from what is in the warehouse. Orders are missed or entered twice. The owner cannot see margins by product or by customer without hours of manual work.
It is also the main reason AI tools struggle in these businesses. A tool that forecasts demand or suggests reorder quantities needs a complete history of orders, stock movements and payments. When that history is spread across chats and files, the tool has nothing reliable to learn from.
The fix usually starts with one stock and order system that the whole team uses, connected to the accounting records. For a growing business, that system is often ERP software sized for an SME. Our guides on when an SME has outgrown spreadsheets and how to choose ERP software for SMEs explain how to decide. For readers who prefer Bahasa Indonesia, our guide on aplikasi stok barang explains how to choose a stock system and when to move to ERP software.
“Stock and orders kept in chats and spreadsheets leave no complete record. Put them in one system first, then add AI.”
Six practical steps
From experimenting with AI to relying on it.
The six steps follow the order in which the foundations are usually built. You can skip ahead, but you will usually have to come back and fix what was skipped, and that tends to cost more than doing it in order.
- 01
Start with a business problem, not a tool.
Name two or three problems where faster work, fewer errors or better insight would improve your results. For distributors and traders, examples include late collections from outlets, stock that does not match the records, and slow order processing. A named problem helps you judge any tool you are offered.
- 02
Map the process the problem sits in.
Write down how the work flows today, who does each step, and where it goes wrong. Include the steps done through messaging apps and paper notes, because these are often where errors start. Our guide to process mapping explains how to do it.
- 03
Check that your systems fit and connect.
Confirm that sales, stock and accounting systems share data without anyone re-entering it. If your business is VAT-registered, check whether tax invoices for Coretax are produced from your own records or keyed in again by hand. Manual steps here usually point to the same gaps that would hold back AI.
- 04
Bring the data together in one trusted place.
Agree how the key figures are defined, clean up the records the AI will use, and store them somewhere the business can rely on. An AI tool can only be as good as the data it reads.
- 05
Know where personal data sits, and who can see it.
List the customer and employee data your business holds, where it is stored, and who has access. Under the Personal Data Protection Law, high-risk processing, which can include AI, may need a data protection impact assessment. Name a senior sponsor who approves what the AI is allowed to do.
- 06
Pilot on a small scale, measure, then expand.
Start the first AI use in one branch, one product line or one team, and measure the results before and after. Expand it only once the improvement holds. A small pilot that works reliably is worth more than a large rollout that later has to be reversed.
Rules and support
Where Indonesia's rules and support fit in.
Indonesia’s rules on data and AI are changing quickly, and support programmes change each year. The table summarises the main points as at October 2026. Always confirm the current position with the relevant ministry or agency before committing any spending.
| Rule or programme | What it covers | Where readiness fits |
|---|---|---|
| Personal Data Protection Law (UU PDP) | Law 27 of 2022 has been fully in force since 17 October 2024. Its implementing regulation, PP 33 of 2026, takes effect on 16 January 2027. It requires breach notification within 3 x 24 hours, impact assessments for high-risk processing, and in some cases a data protection officer. The regulation names AI and machine learning as examples of processing that may need an impact assessment. Fines can reach 2% of annual revenue. | Knowing where personal data is held, and who can access it, is foundation work. It is needed for compliance and for any AI use involving customers or staff. |
| Coretax and e-Faktur | Since January 2025, VAT-registered businesses issue e-Faktur tax invoices through the Directorate General of Taxes' Coretax system. Businesses with annual turnover above Rp4.8 billion must register for VAT. | Tax invoices produced directly from clean system records avoid re-keying. That is the same connected data that AI depends on. |
| Digital Talent Scholarship | The Ministry of Communication and Digital Affairs (Komdigi) runs free digital skills training, including courses on AI. | Training is most useful after processes and systems are settled, so that staff learn the way the business will actually work. |
| Ministry of UMKM programmes | The Ministry of Micro, Small and Medium Enterprises runs programmes to help UMKM adopt digital tools and reach online markets. | Going online is a first step. Using AI in operations also needs defined processes and reliable records behind the sales channel. |
| AI ethics guidance | A 2023 circular from the then Ministry of Communication and Informatics sets out voluntary AI ethics principles. Presidential regulations on an AI roadmap and AI ethics were awaiting signature in September 2026. | The rules for AI are still being written. Clear ownership and good records now will make it easier to meet them. |
Summarised from official sources and announcements. See also the Directorate General of Taxes and Digital Talent Scholarship.
Common pitfalls
Three ways AI projects stall in Indonesian SMEs.
Buying the tool first
A business signs up for an AI tool because a competitor has one or a vendor offered a free trial. Only afterwards does anyone ask which problem it should solve. The tool is then little used, and the next AI proposal becomes harder to approve.
Putting personal data into AI tools unchecked
Staff paste customer lists or chat histories into a public AI tool to save time. Nobody has checked how the provider stores that data. Under the Personal Data Protection Law, the business remains responsible for it.
Records split across chats and spreadsheets
Orders arrive by messaging app, stock is tracked in a spreadsheet, and payments are recorded in the accounting system. No single record shows the whole transaction, so an AI tool has nothing reliable to learn from.
Common questions
Questions Indonesian SMEs ask about AI readiness.
What counts as an SME (UMKM) in Indonesia?
Government Regulation 7 of 2021 defines micro, small and medium businesses by annual sales. A micro business has sales of up to Rp2 billion, a small business up to Rp15 billion, and a medium business up to Rp50 billion. New businesses can instead be classified by their starting capital, excluding land and buildings.
Does the Personal Data Protection Law apply to small businesses?
Yes. Law 27 of 2022 applies to any business that collects or uses personal data, whatever its size. Some duties, such as appointing a data protection officer, apply only to certain types or scales of processing. The implementing regulation, PP 33 of 2026, takes effect on 16 January 2027 and sets out these duties in detail.
Is there a law on AI in Indonesia?
Not yet. A 2023 ministerial circular on AI ethics gives voluntary guidance. Two presidential regulations, one on a national AI roadmap and one on AI ethics, were awaiting signature in September 2026. Personal data rules already apply to AI tools that use personal data.
Do we need Coretax to be ready for AI?
No. Coretax is the tax administration system, and VAT-registered businesses use it to issue e-Faktur tax invoices. It is separate from AI. However, a business that issues tax invoices directly from clean records in its own systems is usually better prepared for AI than one that re-enters invoices by hand.
Where should an Indonesian SME start?
Start with an honest view of where your foundations stand today. Our free Readiness Snapshot takes about five minutes and shows which of the three foundations, processes, systems or data, is most likely to hold back your first use of AI.
Start here
See where your foundations stand.
The free Readiness Snapshot takes about five minutes. It emails you a PDF report showing where your processes, systems and data stand on the path to AI, and which gap to close first.