AI & Automation

Catalytic AI Solutions: a 2026 guide for Singapore manufacturers

By Mark Tan, Founder23 October 202622 min read

Catalytic AI Solutions: a 2026 guide for Singapore manufacturers

On 21 October 2026, at Industrial Transformation Asia-Pacific (ITAP) 2026, Minister of State for Energy, Trade and Industry Gan Siow Huang announced that the government has set aside more than S$45 million over the next three years to help manufacturers adopt AI. The money supports Enterprise Singapore's Catalytic AI Solutions initiative for advanced manufacturing. Eligible companies can get co-funding to adopt AI tools from a suite that EnterpriseSG puts together with trade associations and qualified vendors. One of the tools in that suite automates request for quotation (RFQ) documents.

If you run a precision shop, a contract manufacturer or an equipment maker with 20 to 200 staff, this guide is for you. It sets out what the initiative funds and who it is for, based on what EnterpriseSG and the press have published so far. It also covers what is still unclear and how to prepare before you speak to a trade association. Because most of the manual work we see in manufacturing firms sits around quotations, we spend a full section on preparing an RFQ problem statement, with a timing exercise you can start this week.

The figures below come from two news reports of the announcement, the CNA report and the Straits Times report, both from 21 October 2026.

The programme details come from EnterpriseSG's Catalytic AI Solutions page, checked on 22 October 2026. Where a detail has not been published, we say so. Programme rules can change, so check with EnterpriseSG or its trade association partners before you commit to anything.

What was announced on 21 October

Ms Gan made the announcement at ITAP, held at the Suntec Singapore Convention and Exhibition Centre. The published facts are short:

More than S$45 million has been set aside over the next three years. CNA gives the figure as about US$35 million.
The money supports Catalytic AI Solutions, an EnterpriseSG initiative for advanced manufacturing.
EnterpriseSG will offer SMEs a suite of ready-to-deploy AI solutions and will co-fund eligible companies' adoption of solutions in that suite.
The suite is built with trade associations and chambers and with qualified vendors. ST reports that some firms have already confirmed their interest.
EnterpriseSG says the suite will grow over time as more industry needs and solutions are identified.
Versions for the logistics and wholesale trade sectors are being developed, with details to come later.

"We want AI to be accessible not just to large manufacturers, but also to our SMEs," Ms Gan said at the event, according to the Straits Times.

The initiative is part of Singapore's National AI Missions, which cover advanced manufacturing, connectivity, finance and healthcare. Prime Minister Lawrence Wong announced those missions at Budget 2026. ST describes the advanced manufacturing mission as an effort to reimagine the industry to grow its output and productivity. Manufacturing is around 20% of Singapore's GDP and employs about 12% of the workforce, according to ST.

The same speech covered other news that sits next to Catalytic AI Solutions but is separate from it. The AI Centre of Excellence for Manufacturing (AIMfg), set up by A*STAR and the Ministry of Energy, Trade and Industry in September 2024, is entering its next phase. Both reports say it has worked with 30 manufacturers so far and completed 23 projects that deployed AI. Ms Gan also launched the Singapore Embodied AI Challenge for Manufacturing, which starts with a semiconductor use case from Micron, and the Singapore Manufacturing Federation announced a manufacturing jobs portal. None of these are part of the co-funded suite.

What Catalytic AI Solutions funds

EnterpriseSG's programme page describes Catalytic AI Solutions as an initiative that helps Singapore SMEs "adopt ready-to-use AI solutions, without having to build them from scratch". It brings together solution providers, system integrators and trade associations to develop AI tools for specific sectors. Eligible companies can receive co-funding for the adoption of selected AI solutions.

So the funding is tied to solutions in the suite. It is not described as a general AI budget you can spend on any project you choose. For manufacturing, the programme page lists five solution areas, developed with trade associations and chambers (TACs) and qualified solution vendors:

Request for Quotation (RFQ) Automation: streamline the preparation and processing of RFQ documents
Production Scheduling: anticipate equipment issues and optimise production workflows
Predictive Maintenance and Energy Optimisation: monitor and reduce energy use through real-time analytics, and anticipate equipment failure
Quality Control and Machine Process Improvement: improve inspection processes, efficiency, precision and productivity
Product and Component Design: speed up design cycles and optimise machining parameters with AI-assisted software

CNA reports that EnterpriseSG's media release expects these solutions to cut manual work and response times, improve equipment reliability and keep product quality consistent. Those are the agency's stated aims. None of the three sources reports results from firms that have used the suite.

The initiative is aimed at three barriers that EnterpriseSG names: the cost of developing AI, limited access to solutions built for a specific sector, and uncertainty over outcomes. Its answer is a set of tools that are already proven, so an SME does not have to develop one itself. CNA quotes the agency as saying this could cut the time, cost and complexity of adopting AI.

What has not been published yet

As of 22 October 2026, none of the three sources gives these details:

The co-funding percentage or the cap per company.
Which costs count. For example, whether licence fees, setup, integration, training or the first year of subscription are covered.
How the claim works, and whether you pay first and claim later.
The application form, and the dates it opens and closes.
The names of the qualified vendors behind each solution.
How Catalytic AI Solutions works alongside other grants you may already hold.

One programme partner has published a support level. The Singapore Manufacturing Federation's Catalytic AI page says companies can implement the selected solution with up to 70% funding support, for SMEs and large local enterprises in advanced manufacturing.

Ask EnterpriseSG, or one of the trade associations named below, for each of these before you sign a quotation or pay a deposit. Treat any figure you see elsewhere as unconfirmed until one of them gives it to you in writing.

Who is eligible

The programme page says Catalytic AI Solutions "is open to eligible Singapore-based SMEs, looking to address operational gaps through AI adoption". ST describes the money as being for SMEs in the advanced manufacturing industry. CNA says EnterpriseSG will co-fund "eligible companies' adoption of AI solutions in the suite".

That is as far as the published criteria go. The sources do not set out the SME definition used, any local shareholding rule, minimum years of operation, or which manufacturing subsectors are in scope. The trade associations below can tell you whether your firm qualifies.

What is clear is the route in. The programme page says to start by speaking to a TAC partner, who will help you identify operational gaps, define your problem statement and find the right AI solution provider. For advanced manufacturing, the two partners listed are:

Singapore Manufacturing Federation (SMF)
Singapore Precision Engineering and Technology Association (SPETA)

Logistics and trade firms will have to wait. The programme page says "more information coming soon" for that track, and CNA says details will be announced later.

A quick self-check before you call

You are likely to have a useful first conversation with SMF or SPETA if most of these are true:

You are a Singapore-based manufacturer and you think you are an SME. Confirm your status with the TAC.
You can name one operational gap in a sentence, such as "our RFQs wait too long for an estimator".
That gap matches one of the five solution areas above.
You have recent records of how the work runs today: emails, quote files and job numbers.
A named person in your firm will own the project after the tool goes live.

If you cannot tick the second or fourth point yet, the next section is where to start.

How to prepare before you speak to a TAC partner

The programme page asks you to define a problem statement. A vague one, like "we want to use AI for quotations", gives the TAC partner and any provider little to work with. A good one says which job, how long it takes now, where the time goes and who does it. You can write that from your own records in a week or two, at no cost, before anyone sells you anything.

We suggest four steps.

Step 1. Pick one workflow

Choose the job that costs your team the most hours or the most lost orders. For many job shops and precision firms that is quoting. RFQs arrive with drawings, an estimator reads every title block, someone chases material and outside process prices, and a manager checks the margin before the quote goes out. Scheduling, maintenance, inspection and design are all in the suite too. Pick one area for now and leave the others for later.

Step 2. Time your last 20 RFQs end to end

This is the most useful piece of preparation you can do, and it needs nothing but your inbox and quote files. Pull your last 20 RFQs from the sales inbox or quote log. Twenty is enough to show a pattern and small enough to finish by hand. Include the ones you lost or declined, because slow quotes are often why they were lost.

For each RFQ, record the time stamps you can find in email and file history:

RFQ ID: your quote number or the customer's RFQ number
Received: date and time the RFQ landed in the inbox or portal
First opened: when someone first acted on it (forwarded, logged, replied)
Logged: when it entered your quote tracker or ERP
Clarification out and back: any questions sent to the customer, and when the answer came
Supplier prices out and back: when you asked for material, plating or heat treatment prices, and when they came back
Estimate done: when the estimator finished cycle times and costing
Approved: when the manager or director signed off price and lead time
Sent: when the quote reached the customer
Outcome: won, lost, no reply, or declined

Then work out two numbers for each RFQ: total elapsed time from received to sent, and the hands-on time people spent on it. The gap between them is waiting time: in an inbox, for a supplier, or for a signature. Automation tends to help waiting time and retyping. It does not shorten an estimator's engineering judgement, and it should not try to.

Do not guess the times. If you cannot find a time stamp, leave the cell blank and note why. A blank tells you something too, because it usually means the step lives in someone's head or on WhatsApp.

Step 3. Mark every place data is typed twice

Go through the same 20 RFQs and highlight each time the same piece of information is keyed in by hand more than once. In manufacturing quotes the usual suspects are:

Customer name, contact and RFQ number, copied from the email into the tracker, then into the quote template, then into the ERP.
Part number, revision, material, finish and quantity, read off the drawing and typed into the costing sheet.
Supplier prices, copied from a PDF or email into the costing sheet.
The final price and lead time, typed again into the quote PDF and again into the cover email.
Won quotes, retyped into a sales order or job traveller when the PO arrives.

Count the retypes per RFQ. Note which ones caused an error you can trace, such as a wrong revision or a quantity typo that reached the customer. That list is the most concrete thing you can bring to a TAC partner. It also tells a provider exactly which fields their tool has to read and where they have to land.

Step 4. Note who touches each step

Add one more column to your timing sheet: the name or role of each person who handled the RFQ at each step. Sales coordinator, estimator, purchaser, production planner, manager, director. You will usually find that one or two people sit on almost every quote, and that the waits pile up in front of them.

This matters for any AI tool, funded or not. You need to decide who will check what the tool reads and drafts, and who keeps the final say on price, lead time and what you promise the customer. In every build we do, those decisions stay with a named person in your firm.

Turn it into a problem statement

With the sheet done, you can write a problem statement in plain words. Here is a template. Fill the blanks from your own records and do not round them up.

Over our last 20 RFQs, the median time from received to sent was ___ working days, of which about ___ hours was hands-on work. The rest was waiting for ___. Each RFQ had ___ fields typed by hand ___ times on average, mainly customer details, drawing data and supplier prices. ___ and ___ handle almost every quote. We want to cut the waiting and retyping while our estimator keeps control of cycle times and price.

Take that paragraph and your timing sheet to SMF or SPETA. It gives them what they need to judge whether RFQ automation, or another solution in the suite, fits your gap.

Where RFQ and quotation automation fits

RFQ Automation is the one solution in the suite that sits in the office rather than on the shop floor. The programme page describes it as streamlining "the preparation and processing of Request for Quotation documents". CNA describes it as solutions that "automate the processing of request for quotation documents". Neither source says how the tool works, which systems it connects to, or who supplies it. Ask the TAC partner and the provider for those details.

From the quoting workflows we map with Singapore firms, this is roughly where an RFQ tool can help and where people stay in charge. Use it as a list of questions for any provider, inside the suite or outside it.

The steps a tool can take on

Reading incoming RFQ emails and attachments, and logging each one with a number, an owner and a due date.
Pulling customer details and drawing data such as part number, revision, material, finish and quantity into your costing sheet or ERP, so nobody types them twice.
Matching a new part to similar past jobs, so the estimator starts from history instead of a blank sheet.
Sending supplier price requests and logging the replies against the RFQ.
Drafting the quote in your own template, then the cover email, for a person to check.
Reminding the owner when an RFQ has waited too long, and chasing the customer after the quote goes out.

On our quotation automation page we put it simply: the first win is usually visibility, so that every quote has an owner, a next action and a follow-up date.

The decisions that stay with your people

Cycle times, set-ups and machining strategy. A tool can read a title block. It cannot work out how your shop will make the part.
Price, margin and lead time on every quote.
Whether to quote at all, for example on a part that does not suit your machines.
Releasing the quote to the customer. Nothing should go out on its own.
Anything the customer's NDA restricts. Check where drawings are processed before you connect a tool to your inbox.

Our page on AI automation for precision engineering firms goes into the drawing, traveller and certificate side of this in more detail.

What happens after the quote

An RFQ tool covers the front of the process. The work after it often costs as much time: approving the quote internally, matching the customer's PO to the accepted quote, raising the sales order and the job traveller, and following up on quotes that went quiet. Ask whether the suite solution you are offered covers those steps or stops at the quote. If it stops, plan the handoff now. Our quote-to-PO approval guide sets out a controlled version of that chain, with named owners and no approval by silence.

How Lynqra works with manufacturers on this

Lynqra builds AI agents, automations and small apps for Singapore SMEs, including precision engineering and fabrication firms, and trains the team that runs them. We are not claiming to be a Catalytic AI Solutions provider. For the current list of solutions and providers in the suite, ask SMF or SPETA.

Before you apply, we can sit with your team for the 20 RFQ timing exercise, map who touches each step and draft the problem statement with you. That work is yours to take to a TAC partner, whoever you end up choosing.

If you adopt a tool through the initiative, the steps around it still need to run: internal approval, matching the PO to the quote, follow-up and filing against the job. We can build those around the tools your team already uses, where you and the provider agree on the setup.

If your gap is not covered, or you turn out not to be eligible, we can scope a fixed-price build for that one workflow. Other funding routes, such as EDGE, may be assessed case by case, subject to eligibility and approval. We work with advisory partners on funding and never promise a grant.

In each case a person in your firm keeps the final say on prices, lead times and what goes to customers.

Mistakes to avoid

Paying before you have the rules. The co-funding terms, claim process and dates have not been published in the sources above. If you sign a quotation or pay a deposit before you know them, you may find the cost falls outside what can be funded. Ask first.

Starting with the tool. The programme starts with your operational gap and problem statement, and a provider's demo cannot tell you either of those.

Skipping the baseline. Without your 20 RFQ baseline, you will not be able to show that anything changed. Time your quotes before the tool arrives and again a few months after.

Letting the tool send quotes. Keep a person on price and release, at least until the tool has proven itself on your own RFQs.

Leaving out the people who quote. Your estimator and sales coordinator will run the new flow. Involve them in the timing exercise and in choosing the tool.

Other support mentioned alongside the announcement

CNA notes measures from Budget 2026 that sit beside Catalytic AI Solutions. The Enterprise Innovation Scheme is being expanded to cover qualifying AI expenditure, and the Productivity Solutions Grant will cover a wider range of digital and AI-enabled solutions. Budget 2026 also introduced a Champions of AI programme for companies using AI to transform their business. None of the sources explains how these interact with Catalytic AI Solutions, so check with EnterpriseSG before planning to combine them.

CNA also reports that nearly one in four Singapore businesses adopted AI in 2025, up from about one in seven a year earlier.

For a plain walk through EDGE, the grant that brought EnterpriseSG's business upgrading support together on 30 September 2026, see our EDGE grant guide for SMEs.

Frequently asked questions

What is Catalytic AI Solutions?

It is an Enterprise Singapore initiative that helps Singapore SMEs adopt ready-to-use AI solutions without building them from scratch. EnterpriseSG puts together a suite of solutions with trade associations and qualified vendors, and eligible companies can get co-funding to adopt selected solutions from that suite.

How much money is involved?

More than S$45 million over the next three years, announced on 21 October 2026 for the advanced manufacturing track.

How much of my project will be co-funded?

SMF says on its Catalytic AI page that support is up to 70% for SMEs and large local enterprises in advanced manufacturing. EnterpriseSG's programme page and the CNA and Straits Times reports do not give a figure or a cap, so confirm what applies to your project with SMF, SPETA or EnterpriseSG.

Is RFQ automation covered?

Yes. Request for Quotation (RFQ) Automation is one of the five manufacturing solutions on EnterpriseSG's programme page. The others are production scheduling, predictive maintenance and energy optimisation, quality control and machine process improvement, and product and component design.

Who can apply?

The programme page says it is open to eligible Singapore-based SMEs looking to address operational gaps through AI adoption. The detailed criteria have not been published in the sources we checked, so confirm your eligibility with SMF or SPETA.

How do I get started?

Speak to a TAC partner. For advanced manufacturing, the programme page names the Singapore Manufacturing Federation (SMF) and the Singapore Precision Engineering and Technology Association (SPETA). They help you identify your operational gaps, define a problem statement and find a suitable provider.

Can I use any AI vendor I like?

The co-funding is for selected solutions in EnterpriseSG's suite, developed with qualified vendors. Check with the TAC partner which providers and tools qualify before you commit.

Is Lynqra an approved provider under the initiative?

We are not claiming to be. Ask SMF or SPETA for the current list. We can help you prepare your problem statement and baseline, build the workflow around a suite solution, or build a separate workflow if the suite does not fit.

We are a logistics or wholesale trade company. Can we join?

Not yet. EnterpriseSG is developing versions for logistics and wholesale trade, and says details will be announced later.

What should I prepare before the first meeting?

A one-paragraph problem statement and the records behind it. For quoting, time your last 20 RFQs from received to sent, mark every place data is typed twice, and note who touches each step.

Related guides

Bring your last 20 RFQs

If you want help with the timing exercise or the problem statement, book a free consult. Bring a recent RFQ with its drawings and your quote log. We will map how quoting runs today and tell you plainly where a tool would help and where it would not.

Sources

← Back to Blog