UKRI announced on 10 September 2026 that it will change how grant applications are assessed. The aim is to cut unnecessary work, speed up decisions and respond to generative AI without weakening the quality of funding decisions.
Read the assessment route before drafting
UKRI’s five-year strategy targets a reduction of at least 50% in grant-processing times by 2031. For applicants, the immediate change begins with the funding opportunity itself: from October 2026, UKRI plans to combine eight measures, but not every opportunity will use the same combination.
Each call will say whether demand management applies and which approach is being used. UKRI describes three levels: standard measures, additional measures and application caps. Additional measures may include an expression of interest, a preliminary application or a preselection process. Caps may be proportional or fixed.
Use that information to set the order of work. If an expression of interest comes first, treat it as a genuine decision point rather than preparing the full proposal immediately. If the opportunity has a cap or preselection stage, confirm the limit and screening route before investing substantial time. UKRI says it will use demand management only when necessary and select the least intrusive tool that preserves fairness, quality and public value. Read UKRI’s explanation of demand management
The practical benefit is a clearer early checkpoint. The compromise is that some applicants may need to demonstrate fit before reaching a full peer-review stage.
Let the published questions shape the proposal
UKRI plans to standardise and simplify processes across its research councils, introduce consistent question and assessment-criteria sets, and remove unnecessary attachments from its Funding Service. An older application habit therefore may not transfer cleanly to a new call.
Start with the questions and criteria published for that opportunity. Map each answer to the criterion it addresses, then remove material that does not help an assessor judge it. Additional documents are not automatically an advantage: UKRI’s stated direction is to remove attachments that are not needed, while retaining anything the individual opportunity still requests.
This approach also makes the application easier to check before submission. Every required answer should have a clear purpose, and every attachment should exist because the call asks for it or because it directly supports an assessment requirement.
Expect the review method to vary by opportunity
Programme Directors and Opportunity Leads will be able to choose assessment arrangements suited to an opportunity’s objectives, expected outcomes and scale. UKRI intends to increase the use of Distributed Peer Review for high-volume opportunities, building in part on trials run by the ESRC and BBSRC.
UKRI also plans to test initial staff triage, including staff-led rejection based on the written application, with support from the UK Metascience Unit. Where a call uses screening before wider review, put the project’s purpose, intended outcome and fit with the criteria where they can be found quickly. Keep the supporting evidence precise and relevant; a longer application is not automatically stronger when the route is designed to manage high demand.
Another planned experiment is weighted multi-criteria assessment. This is intended to balance excellence with the strategic objectives of each opportunity. Scientific or technical quality remains important, but it may not be the only dimension carrying weight in a particular funding route. Check the weighting and criteria for that call instead of applying one template to every UKRI application.
Handle AI use according to the applicable rules
UKRI says generative AI can assist with drafting and assessing applications, while raising concerns about confidentiality and the integrity of peer review. The organisation is exploring AI-assisted assessment with controls and safeguards. Its first planned experiment is retrospective: with the UK Metascience Unit, UKRI will compare what an AI system would have concluded from older applications with the decisions that were actually made.
Before using generative AI on application material, check the instructions for the relevant opportunity and your organisation’s research office. Do not treat AI-generated text as evidence, and review every factual, methodological and attribution claim yourself. UKRI’s announcement does not establish one universal permission or prohibition for applicants, so assumptions about acceptable tools or confidential material should not replace the applicable guidance.
UKRI says human judgement will remain at the centre of funding decisions. It also plans to monitor implications for equality, diversity and inclusion and work with applicants, research organisations and their research services. AI may become part of the wider assessment system, but the announced direction does not make automated decisions the default.
Prepare for extra assurance on higher-risk funding
For high-risk funding, UKRI plans in-person verification, such as an interview, when additional assurance is needed. If an opportunity includes this step, the people responsible for the proposal should be able to explain its objectives, expected outcomes and underlying evidence consistently.
An interview or other verification step is an additional assurance route, not a stated requirement for every application. Its presence should affect who is involved in preparation and how clearly the proposal’s evidence and responsibilities are documented.
Use this order before submitting
First, confirm whether demand management applies. Next, identify any expression of interest, preliminary stage, preselection process or application cap, including whether a cap is proportional or fixed. Then map every answer to the published questions and criteria, remove unrequested material, and check any weighting applied to the assessment.
After that, verify how the opportunity treats generative AI and whether your research office has additional instructions. Finally, check for extra assurance requirements such as an interview and make sure the relevant people can support the proposal if that stage is used.
UKRI says no single measure suits every opportunity. Several measures may be combined, while others will be tested and introduced as evidence accumulates. The safest preparation is therefore opportunity-specific: read the route first, then shape the proposal around its questions, criteria and screening steps.
The planned changes could shorten processing times and reduce avoidable work, but they may also move selectivity earlier through caps, triage or weighted criteria. Applicants should prepare for a process that is more clearly tailored to each call, rather than assume that one universal UKRI application route will apply from October 2026.
