---
title: AI grants UK small businesses can actually claim in 2026 (Innovate UK, Made Smarter, BridgeAI decoded)
description: Most UK small businesses skip the AI grants they qualify for because the application takes 40 hours and the money is post-spend. After helping clients through Innovate UK Smart Grants, Made Smarter assessments, and BridgeAI matched funding, here is the practitioner version. Which programmes actually pay out, what they expect, where the time-traps are, and the four-week plan to apply without losing a quarter of your year to a 50-page form.
canonical: https://richardbatt.com/blog/uk-ai-grants-2026-innovate-made-smarter
date: 2026-05-05
author: Richard Batt
tags: [UK Grants, Innovate UK, Made Smarter, AI Funding]
type: blog_post
---

# AI grants UK small businesses can actually claim in 2026 (Innovate UK, Made Smarter, BridgeAI decoded)

_Most UK small businesses skip the AI grants they qualify for because the application takes 40 hours and the money is post-spend. After helping clients through Innovate UK Smart Grants, Made Smarter assessments, and BridgeAI matched funding, here is the practitioner version. Which programmes actually pay out, what they expect, where the time-traps are, and the four-week plan to apply without losing a quarter of your year to a 50-page form._

**Richard Batt** — AI implementation specialist. 120+ projects across 15+ industries, serving SMBs (5-200 employees) worldwide from Middlesbrough, UK (working globally). Contact: richard@richardbatt.com · https://richardbatt.com

Most UK small business owners I meet have heard the word "Innovate UK." A few have heard of Made Smarter. Almost none can tell me what BridgeAI is. So when the Treasury announces another AI funding pot, the press release lands, and a year later the same firms tell me they didn't apply because the form looked terrifying and the deadline came and went.

That's the practitioner reality of UK AI grants in 2026. The money is real. Tens of millions of pounds, region by region, sector by sector. The problem isn't that small firms can't qualify. The problem is that the application is a project of its own, and a 30-person company has nobody whose job it is to write 40 pages of strategic narrative on a Wednesday afternoon.

After 120 AI projects across 15 industries, I've watched clients win grants and watched clients quietly drop out at week three. This post is the practitioner version. Which UK programmes are live in 2026, what they actually fund, the eligibility traps, and the four-week plan to apply without losing a quarter of your year to a form.

**TL;DR**

- Innovate UK Smart Grants fund collaborative AI projects from £25,000 to £500,000, but the bar is genuine novelty plus a strong consortium.
- Made Smarter (regional, manufacturer-only) covers up to half the cost of digital adoption projects from £20,000 to £100,000, with grants typically £5,000 to £50,000.
- BridgeAI, run by Innovate UK Business Connect, focuses on AI uptake in agriculture, construction, creative, and transport. It's the most accessible programme for non-technical SMBs.
- R&D tax credits are the quiet workhorse. They reimburse 25 to 33 percent of qualifying AI development spend after the year ends, with no application gauntlet.
- Allow 40 hours of senior time across four weeks for any serious application, plus a buy-in conversation with whoever runs your finance.

## What counts as an AI grant in 2026

Define the term, because the press uses it loosely. A UK AI grant is a non-repayable cash award from public funds (national or regional) given to a business so it can adopt, adapt, or build artificial intelligence capability. Some grants reimburse spend after the project finishes. Some give cash up front. None will fund a tool you would have bought anyway.

There are four broad categories live in 2026 that a small or mid-sized business (SMB) should know about: national R&D funding (Innovate UK), regional digital adoption funding (Made Smarter), sector-specific uptake funding (BridgeAI), and tax-side support (R&D tax credits, Patent Box). Most owners only know the first one. The other three pay out more often.

## Smart grants, the headline route from Innovate UK

Smart Grants is the headline programme. It's funded by UK Research, run through Innovate UK, and it backs projects that develop genuinely novel products or processes. AI features heavily because most novel work in 2026 has a model in it somewhere.

The numbers: grants of £25,000 to £500,000 for single-company projects, up to £2 million for consortia. Funding intensity is 70 percent for small companies on feasibility work and 60 percent for industrial research. You match the rest from your own cash or other private sources.

Eligibility looks open on paper. In practice, the bar is real novelty. If your project is "we'll install ChatGPT and our staff will use it for emails," don't apply. But if your project is "we'll build a model that predicts maintenance failures in lithium battery packs and license the algorithm to the wider industry," that's a Smart Grant.

The application is the trap. You need a clear technical narrative, a costed work plan, a market analysis, a route to commercialisation, and an impact statement. Allow 60 to 80 hours of senior time. The win rate is around 8 to 12 percent in any given round, so write the proposal once and re-use the structure if you miss the first cut.

A 22-person engineering firm I worked with last quarter pulled in £180,000 of Smart Grant funding to build a defect-detection model for their CNC line. They didn't write the application alone. They paid a specialist proposal writer £6,000 and used my technical content to shape the AI section. The proposal landed. The project is now in month four of nine.

## Made Smarter (regional, manufacturer-only)

Made Smarter is the practical workhorse for makers. It started in the North-West, then expanded to Yorkshire and the Humber, the East Midlands, the West Midlands, the North-East and Tees Valley, and the South-West. Each region runs its own version with its own grant ceiling, but the model is consistent: free advice plus matched funding for digital projects.

Eligibility is narrow and worth checking. You need to be a UK manufacturing SMB (typically under 250 employees), based in the region, with the ambition to adopt digital technology that will measurably improve your operations. Pure services firms, retailers, and most professional service businesses won't qualify, however much AI they want to buy.

The numbers: matched funding usually covers 50 percent of project costs, with grants from £5,000 up to £50,000 depending on region. Total project size sits in the £20,000 to £100,000 band. And the regional team gives you a free digital readiness review first, which itself is worth a day of consulting time.

Made Smarter will fund AI adoption work that has clear operational outcomes: vision systems on a packing line, a forecasting model wired into ERP (enterprise resource planning), an automated quality inspection workflow. But it won't fund pure research, training-only spend, or off-the-shelf software you could have bought without the funding.

A North-East food manufacturer I supported in late 2025 used Made Smarter Tees Valley to part-fund a £40,000 project. We installed a vision-based reject system on their inspection line, plus a forecasting model on their order book. The grant covered £18,000. Their own match was £22,000. Payback on the operational side came inside seven months.

## BridgeAI (sector uptake, the accessible one)

BridgeAI is the youngest of the four and the most useful for non-technical small firms. Run by Innovate UK Business Connect (the network arm of UKRI), it focuses on AI uptake in four sectors: agriculture and food, construction, creative industries, and transport. The premise is that these sectors lag in AI adoption, and the cure is supported pilots rather than capital injection.

The numbers move around between calls, but the typical shape: technical assistance vouchers worth £15,000 to £45,000, plus competition rounds for matched-funded pilots up to £1 million for consortia. The voucher route is the most practical for an SMB. You spend the voucher with a vetted technical partner. They build the pilot. You keep the output.

Eligibility is straightforward. You need to be a UK-registered SMB in one of the four target sectors, with a defined business problem AI could plausibly help with. You don't need a technical co-founder or a research department. The team at Business Connect will help you scope the problem before you apply.

The trap with BridgeAI is timing. Calls open and close on schedules that don't always match your operational year. So if you miss the autumn call you may be waiting until spring. Plan ahead. Sign up for the Business Connect newsletter and watch the call calendar.

A 14-person construction-services firm I worked with used a BridgeAI voucher in 2025 to build a quote-generation model from their ten-year project archive. The grant covered £22,000 of the £30,000 build. The model now drafts an initial bid in 40 minutes for a job that used to take a senior estimator a full day.

## R&D tax credits (the quiet workhorse)

R&D tax credits aren't a grant. They're a tax-side reimbursement, and that distinction matters because nobody has to win them in a competitive process. If your business spent money in the last accounting period on developing a product or process that resolved a real technical uncertainty, you can claim back a portion of that spend after the year ends.

The numbers in 2026: the merged R&D scheme reimburses around 16-20 percent of qualifying spend for most companies, rising to 27 percent for R&D-intensive SMBs (those spending more than 30 percent of their cost base on R&D). Loss-making intensive firms get a higher cash credit rate.

Eligibility for AI work is generous if the work meets the technical bar. Building a model that didn't exist before, integrating a model into a non-trivial pipeline, or solving a data problem that public guidance doesn't answer can all qualify. But buying an off-the-shelf tool and using it does not.

The application is a tax filing rather than a grant proposal. You write a technical narrative for HMRC, your accountant lodges the claim with your corporation tax return, and the credit lands as a reduction in your tax bill or a cash payment. Allow 15 to 20 hours of senior time, plus an external R&D tax adviser fee of typically £2,000 to £6,000 (often contingent on success).

The watchout: HMRC has tightened R&D enforcement since 2024. If the technical narrative is thin, claims now get rejected at audit. So use a specialist adviser. Avoid the firms that promise "guaranteed claim" on a no-questions basis. A 30-person services firm I worked with last month claimed £42,000 in R&D credits for AI development work across their previous financial year. The narrative was tight because the project was real. That matters.

## How the four programmes compare

| Programme | Typical grant size | Match required | Sectors | Time to apply | Speed of decision |
| --- | --- | --- | --- | --- | --- |
| Innovate UK Smart Grants | £25k to £500k | 30-40% | All sectors | 60-80 hours | 3-4 months |
| Made Smarter | £5k to £50k | 50% | Manufacturing only, regional | 15-25 hours | 6-10 weeks |
| BridgeAI vouchers | £15k to £45k | 0-30% | Agri, construction, creative, transport | 10-20 hours | 4-8 weeks |
| R&D tax credits | 16-27% of spend | None | All sectors | 15-20 hours | After year-end |

If you only have time for one application this year, the order of practical priority for most SMBs is: R&D tax credits first (low effort, high hit rate), then Made Smarter or BridgeAI if you qualify by sector, then Smart Grants only if you have a genuinely novel project and a partner willing to write with you.

## The four-week application plan

Most owners I work with don't have 80 hours to spend on a Smart Grant alone. Here's the plan that works for the smaller programmes and that scales up if you commit to a Smart Grant later.

**Week one: scoping.** Pick one operational problem AI could plausibly solve. Write it out in 200 words. Quantify the current cost in hours per week, mistakes per month, or revenue at risk. This becomes your problem statement and you'll reuse it in every application.

**Week two: programme fit.** Read the eligibility criteria for the three programmes that match your sector. Book the free-readiness call with Made Smarter or Business Connect. Ask the assigned advisor which programme fits your problem best and what they have seen succeed in your sector. Their answers save weeks.

**Week three: drafting.** Write the technical narrative, the work plan, and the budget. Keep the narrative concrete. Numbers, not adjectives. If the form asks for impact, give specific operational metrics and a plausible payback period. Use your accountant for the financial sections. Use your operations manager for the work plan.

**Week four: review and submit.** Have one external person read the draft cold. They'll spot the parts that read as buzzword soup. Tighten, then submit. Save every draft. The same scaffolding will speed up the next application by a factor of three.

The whole exercise costs around 25 to 40 hours of senior time. That's a real cost, but it isn't a quarter of your year. The reason most owners drop out is that they think they need 80 hours up front. They don't, unless they're chasing Smart Grants on day one.

## Where SMBs lose the most money

Three patterns I see week after week.

Pattern one: applying to the wrong programme. A bookkeeping firm I spoke to in March wanted Smart Grant money for "we'll use AI in our practice." That isn't a Smart Grant project. It's an operational change. The right programme for them was R&D tax credits on the integration work plus a small BridgeAI voucher. They wasted six weeks on the wrong form.

Pattern two: post-spend surprise. Most grants reimburse after you spend the money. If you don't have the cash flow to fund the project up front and then claim later, the grant won't help you. So talk to your accountant before you apply, not after. Some grant managers will release tranche payments against milestones, but most won't.

Pattern three: ignoring R&D tax credits. The credit pays out every year you spend on qualifying work, with no application competition. And yet half the SMBs I meet have never claimed. Their accountant hasn't flagged it because their accountant isn't specialised in R&D. A specialist adviser, paid contingent, is worth the call.

## How my clients use the funding

I work with grant-aligned projects regularly. The pattern that holds is to design the project for operational return first, then check whether a grant fits. Never the other way around. Designing for the grant gives you a project that wins the form and loses the operational case.

In practice this means I scope the AI workflow with the client, build the operational business case (hours saved, errors reduced, revenue protected), then look at whether Made Smarter or BridgeAI maps onto the project. If the grant fits, we shape the proposal around the work that was already going to happen. But if it doesn't, we proceed without the grant rather than redesigning for it.

The AI Roadmap audit at richardbatt.co.uk/roadmap is built to surface those operational candidates first, then map them to programmes that will actually pay out. If you want a structured starting point that gives you the project before the form, the Roadmap is the fastest route.

## Frequently asked questions

**Can I claim multiple AI grants at the same time?**

Sometimes, but rarely on the same project. Most schemes have anti-stacking rules that stop you funding the same costs twice. You can run different programmes on different projects in parallel without issue. R&D tax credits are usually compatible with grants, but with rules: the qualifying spend reduces by the grant amount.

**How long does the money take to arrive?**

Decisions on smaller programmes (Made Smarter, BridgeAI vouchers) take 4 to 10 weeks. Smart Grant decisions take 3 to 4 months. R&D tax credits arrive after your accounting year ends, usually 6 to 12 weeks after HMRC accepts the claim. Plan cash flow around the slow timelines, not the press release.

**Do I need a consultant to apply?**

For Smart Grants over £100,000, almost always. Specialist proposal writers raise win rates from low single digits to 25 percent or better. For Made Smarter and BridgeAI, no. The regional teams help you draft. But for R&D tax credits, use a specialist adviser, not your general accountant.

**What if my project isn't novel enough for Smart Grants?**

Apply to Made Smarter or BridgeAI instead. They reward adoption, not invention. A clear operational case for putting AI into your workflow is enough. So save the Smart Grant for the project that actually pushes a technical boundary.

**Can a services firm get any of this money?**

Yes, but the route is narrower. Smart Grants are open to all sectors. BridgeAI covers creative industries (a wide tent that includes most digital agencies). R&D tax credits are sector-blind. Made Smarter is manufacturer-only and most professional services firms won't qualify there.

**What is the single biggest mistake on grant applications?**

Vague impact claims. "This will transform our operations" loses every time. "This will reduce quote-generation time from 6 hours to 40 minutes per bid, across an estimated 200 bids per year, saving 1,067 hours of senior estimator time" wins. Numbers, not adjectives.

## What to do this week

Pick one of two starting points. If you've got an operational problem that AI could plausibly help with, sketch the 200-word problem statement and book the free readiness call with Made Smarter or BridgeAI in your region. But if you've already done AI work in the last accounting year, talk to a specialist R&D tax adviser this fortnight, not after the year ends.

If you want a structured approach that puts the operational case first and matches it to programmes, the AI Roadmap audit at richardbatt.co.uk/roadmap is built for exactly this. The AI Ops Vault at richardbatt.co.uk/vault has the proposal templates and the workflow scaffolds I use with clients during application weeks.

After 120 AI projects across 15 industries, the pattern is the same. The grant money is real. The form is finite. And the firms who apply consistently win more often than the ones who wait for a press release that says it's got easier. It won't.

---

## More about Richard Batt

Richard Batt is an AI implementation specialist who helps businesses deploy working AI automation in days, not months. 120+ projects across 15+ industries.

### Key pages

- [Home](https://richardbatt.com/)
- [About Richard](https://richardbatt.com/about)
- [Blog](https://richardbatt.com/blog)
- [Contact](https://richardbatt.com/contact)
- [Subscribe](https://richardbatt.com/subscribe)

### Contact

- Email: richard@richardbatt.com
- Location: Middlesbrough, UK (working globally)
- Website: https://richardbatt.com