UKRI Enabling AI Adoption Across Science and Engineering 2026: Up to £250,000 for Exploratory AI Research

UKRI Enabling AI Adoption Across Science and Engineering 2026: Up to £250,000 for Exploratory AI Research

The UK Research and Innovation (UKRI) funding opportunity Enabling AI adoption across science and engineering is designed for research teams that see a specific scientific or engineering problem where artificial intelligence could change how the work is done. It is not a general-purpose AI grant and it is not primarily a call for improving an existing model. The application must put an important, application-focused research challenge at the centre, then show why a partnership between AI specialists and domain researchers is needed to investigate it.

The call is particularly relevant to teams working in engineering biology, advanced materials, quantum technologies, medical research, or fusion energy. UKRI describes the activity as ambitious and exploratory: applicants can test existing AI capabilities in new settings, explore requirements for future tools and methods, or build evidence about the barriers to responsible adoption. Projects may involve uncertainty and risk, but they still need a clear question, credible work plan, and useful path to longer-term capability.

The opportunity is currently listed as upcoming. Applications are scheduled to open on 29 September 2026 at 9:00am UK time and close on 1 December 2026 at 4:00pm UK time. The fixed project start date is 26 April 2027, and an award can run for up to 12 months. The call is delivered through the new UKRI Funding Service and uses distributed peer review, so a team applying should plan for both proposal preparation and reviewer participation.

Key details

The £250,000 figure is the maximum full economic cost rather than a promise that every successful project receives £250,000 in UKRI cash. UKRI says it will fund 80% of full economic cost. As a simple illustration, a project budgeted at the maximum full economic cost would imply a standard UKRI contribution of £200,000, with the balance and the treatment of individual cost lines determined by UKRI rules and the host organisation. Use the Funding Service guidance and your research office’s costing process for the actual request.

What the grant is meant to support

UKRI frames this call around a gap between strong AI capability and the practical needs of research disciplines. A domain community may have valuable data, a difficult modelling problem, or an experimental workflow, but not yet know which AI approaches are reliable, useful, or responsible in that setting. Conversely, AI researchers may have methods that have not been tested against the constraints of a real scientific problem. The grant is meant to bring those perspectives together.

The work can include adopting an existing AI method in a novel research context, identifying a need for a new tool or dataset, evaluating how AI could improve discovery, or studying what prevents a research field from using AI effectively. Examples in the engineering-biology scope include modelling cellular and metabolic systems, predictive genome design and engineering, high-throughput phenotyping, and biosecurity or biological risk management. Those examples are directions, not a checklist: a proposal still needs to define its own important challenge and explain why the AI work matters to it.

The call also expects a legacy beyond the immediate experiment. UKRI is interested in skills, collaborations, networks, methods, or resources that can support continued AI adoption after the award. A credible application should therefore describe what the team will leave behind: for example, a validated workflow, a well-documented dataset or benchmark, a reusable analysis method, trained researchers, a tested collaboration model, or evidence that informs a larger future research programme. Do not promise every type of output. Choose the few that follow naturally from the proposed problem and explain how they will be maintained or used.

Who can apply

The project lead and lead organisation must satisfy UKRI’s standard eligibility rules. The official call states that the lead must be based at a UK research organisation eligible for UKRI funding. That makes this a research-organisation-led opportunity rather than a direct grant for an independent startup or an individual working without an eligible host. The host should confirm its status and internal approval requirements before the team invests time in a full application.

The team should include the expertise needed to carry out the plan. UKRI specifically expects teams to reflect the skills and experience required by the proposed activity, including both AI expertise and domain expertise. Joint leadership is allowed, but the team must agree which project lead will be responsible for submitting the application. If the team proposes job sharing, it must explain and justify the arrangement in the applicant-and-team-capability section.

International researchers have a limited route into this call. Because the Engineering and Physical Sciences Research Council is a lead funder, an international researcher may apply as a project co-lead only when the application uses the UKRI-RCN Money Follows Cooperation agreement or the UKRI-IIASA agreement. Other international collaborators, and UK partners who are not based at approved organisations, should be included as project partners where appropriate. Confirm the applicable agreement and organisational position with the lead institution rather than assuming that a normal international co-investigator route is available.

UKRI also says that uninvited resubmissions of projects previously submitted to UKRI or another funder will not be accepted. A revised idea needs to be checked carefully against that rule. Changing the title or adding an AI work package does not automatically turn a resubmission into a new proposal.

Priority areas and fit

Every application must address a domain-specific challenge in one of five areas: engineering biology, advanced materials, quantum technologies, medical research, or fusion energy. The strongest fit is likely to be a proposal where the domain problem is important even before AI is mentioned. AI should provide a reasoned route to new evidence, faster or better analysis, a new experimental capability, or a clearer decision—not simply appear as a fashionable method.

For a good fit, the team should be able to answer four questions early in the proposal:

1. What scientific or engineering challenge is blocking progress?
2. What AI approach or capability will be explored, and why is it plausible for this problem?
3. Why must AI specialists and domain researchers work together rather than handing the task from one group to another?
4. What evidence will show whether the approach works, is responsible, and is worth developing further?

A proposal can be exploratory without being vague. State the uncertainty you are testing, identify the decision points in the work plan, and describe what a useful result would look like even if the first method fails. For medical or engineering-biology projects, explain data governance, safety, privacy, or biosecurity considerations at the level appropriate to the work. For quantum, materials, or fusion projects, connect computational outputs to measurable physical or experimental consequences. UKRI’s aim is adoption across research, so a technically impressive demonstration with no credible user or disciplinary context may be a weak fit.

Funding, duration, and project planning

The call has a £5 million total fund and allows a project full economic cost of up to £250,000. UKRI will fund 80% of the full economic cost. Projects must last no more than 12 months and must use the fixed start date of 26 April 2027. That makes the planning horizon unusually important: a team should define a bounded exploratory programme that can generate meaningful evidence in one year, rather than describing a multi-year research agenda with a small first tranche.

Build the budget around the evidence needed to answer the central question. Depending on the project, that might involve researcher time, access to data or compute, experimental work, software or specialist services, workshops with domain users, or dissemination and documentation. Do not assume that every attractive cost is eligible or that a partner can be paid in the same way as a core team member. The official page directs applicants to the UKRI Funding Service and standard eligibility information; the host research office should translate the work plan into the permitted cost categories.

The fixed start date also affects hiring, procurement, data access, and partner agreements. Identify anything that could delay the first month of work and include a realistic route around it. A proposal that spends half of a 12-month award waiting for approvals will look less credible than one that has made the dependencies explicit and sequenced the exploratory work accordingly.

UKRI lists contributions to project-lead and co-lead salaries, research, technical and project-management staff, research consumables including compute, travel and subsistence, data preservation and sharing, dissemination, estates, and indirect costs among the costs it will fund. It will not fund PhD studentship costs or equipment costing more than £25,000 per item including VAT. Those exclusions are easy to miss when a team is designing a technically ambitious project, so test the proposed work against them before finalising the budget.

Application process and timeline

The application window opens on 29 September 2026 at 9:00am UK time. UKRI expects the application to be made through the new Funding Service, not the Joint Electronic Submissions system. The project lead is responsible for completing the submission, while all team members and project partners are expected to contribute. Only the lead research organisation can submit the application.

The official process is straightforward but should not be left to the final week:

1. Confirm that the host organisation and project lead meet UKRI eligibility rules.
2. Register or sign in to the UKRI Funding Service and select the correct opportunity when it opens.
3. Agree the project lead, administrative lead, AI specialists, domain specialists, partners, and the people who will participate in peer review.
4. Develop the challenge, objectives, method, team capability, costings, responsible-AI approach, and expected outputs together.
5. Answer the Funding Service questions directly and upload any documents requested by the service.
6. Complete the plain-English summary, which UKRI gives a 550-word limit, without including confidential information.
7. Obtain institutional approval and submit before 4:00pm UK time on 1 December 2026.

UKRI has announced a webinar for potential applicants on 14 October 2026 from 10:00am to 11:00am. Use that event to resolve uncertainties about scope, the Funding Service, or distributed peer review, and record any later clarification on the official page. UKRI says the opportunity is subject to Full Business Case approval, so applicants should treat the call as a live upcoming opportunity while continuing to check the official listing for updates.

What reviewers will look for

This call uses distributed peer review (DPR). Applicants agree to review other applications submitted to the same opportunity, and each application receives up to eight reviews that are used to rank applications and make the funding decision. There is no project-lead response or panel step in the process described on the official page. The team must identify at least two people who will take part in DPR, including the project lead, and applicants can be involved in only two projects.

That model makes clarity especially valuable. Write for expert peers who may understand either the AI method or the domain problem, but not both in equal depth. Define specialist terms, make the causal chain from problem to method visible, and state what evidence will change the team’s view. A reviewer should be able to tell what is novel, what is uncertain, what can be completed by April 2028, and why the collaboration is necessary.

Evidence quality matters more than a long list of possible benefits. If the project will compare AI-assisted and existing approaches, specify the evaluation basis before describing the anticipated result. If the project will build a capability, define the users, test conditions, documentation, and adoption barrier. If the project deals with sensitive data or high-consequence decisions, explain the safeguards and limits of the proposed work. Avoid claiming that AI will transform an entire field when the grant can only test one focused research pathway.

Materials to prepare

Start assembling the following before the Funding Service opens:

• A one-paragraph problem statement naming the science or engineering challenge and its practical consequence.
• A short explanation of why the selected AI approach is appropriate and what uncertainty the project will test.
• A work plan for the fixed 26 April 2027 start and the maximum 12-month duration, including dependencies and decision points.
• A team-capability map showing what the AI and domain experts each contribute and how they will work together.
• A costed outline based on the full economic cost, with the 80% UKRI contribution clearly distinguished from other support.
• A plan for data, software, compute, experimental access, ethics, safety, privacy, or biosecurity as relevant.
• Measurable outputs and a route for sharing skills, methods, resources, or collaborations after the award.
• A plain-English summary within the 550-word limit and suitable for public release.
• Confirmation that the selected reviewers have no conflicts and that at least two eligible team members can participate in DPR.

Ask the host research office to review eligibility and costings early. Ask a domain researcher who is not on the team to read the challenge description, and ask an AI specialist to read the domain explanation. If either reader cannot explain the proposed test and its success criteria after one pass, the application probably needs a tighter structure.

Common mistakes to avoid

The most serious mistake is treating the opportunity as a generic AI funding call. A project must sit in one of the five priority areas and centre an application-focused challenge. A second mistake is making the domain partner decorative: the proposal should show how the collaboration changes the question, method, or evidence, not just list two disciplines in the team section.

Do not budget the full £250,000 as if it were guaranteed UKRI cash. It is the maximum full economic cost, and UKRI’s stated contribution is 80%. Do not write a two- or three-year programme around a one-year award. Do not omit the fixed start date. Do not promise a route for international co-leadership unless the application uses one of the specific agreements named by UKRI.

Finally, do not overlook the reviewer obligation. A team that cannot identify appropriate DPR participants or that has one person attached to too many projects could create an avoidable eligibility or process problem. Check the live Funding Service questions and the official page again immediately before submission because the opportunity is still listed as upcoming and UKRI may publish clarifications.

FAQ

Is this open now?

The official status is upcoming. The application window is scheduled to open on 29 September 2026 at 9:00am UK time and close on 1 December 2026 at 4:00pm UK time. Prepare now, but submit only through the live UKRI Funding Service opportunity.

Can a UK startup apply directly?

The call says the project lead must be based at a UK research organisation eligible for UKRI funding. It is therefore not presented as a direct standalone startup competition. A company may be relevant as a project partner where the rules allow, but the lead organisation should confirm the arrangement with UKRI and its research office.

Does UKRI fund the whole £250,000?

No. £250,000 is the maximum full economic cost, and UKRI says it will fund 80% of that cost. The precise request depends on eligible costs and institutional costing. A maximum-cost project would imply £200,000 at 80%, but that is a calculation, not a guaranteed award amount.

Can an international researcher be the project lead?

The official call requires the lead to be based at an eligible UK research organisation. International researchers may participate as project co-leads only under the UKRI-RCN Money Follows Cooperation agreement or the UKRI-IIASA agreement. Other international collaborators should normally be treated as project partners.

What happens after submission?

Applications use distributed peer review. Applicants agree to review other submissions, each application can receive up to eight reviews, and those reviews are used to rank applications and make the funding decision. The call does not describe a project-lead response or panel step.

Is a previous proposal eligible as a starting point?

UKRI states that it will not accept uninvited resubmissions of projects submitted to UKRI or another funder. If your idea has already been submitted, contact the relevant research office or UKRI before treating a changed version as a new application.

Official links and next steps

Read the official UKRI opportunity page for the live application link, full eligibility guidance, Funding Service questions, webinar information, and any updates before submission. The page is the authority for the final scope, cost rules, and deadline.

For a practical next step, identify the exact research challenge, the domain lead, and the AI lead before the 29 September opening date. Then ask the host organisation to confirm eligibility, costing, partner status, and the two-person DPR commitment. A focused question, a credible one-year test, and evidence that the team can continue using what it learns will give the application a much stronger foundation than a broad promise about AI transformation.