What evidence belongs under UKDI's desirable, feasible and viable criteria?

A strong UK Defence Innovation proposal needs three different kinds of evidence. Customer relevance cannot substitute for a credible research method, and a technically impressive method cannot substitute for an executable project. Treat the application as an argument about a specific investment in learning, rather than a general description of the company.

UKDI's published baseline evaluates desirability, feasibility and viability. These cover strategic fit and user need; technical credibility, innovation and team capability; and cost, value and delivery planning. Competition documents can adjust the baseline. The guidance also states that letters of support are not required. Check the precise call before arranging the application around a generic template. UKDI assessment guidance The current DASA-to-UKDI calls guide identifies where to check the current opportunity before preparing an application.

For a software company, desirability starts with the decision its customer struggles to make. “Our platform uses machine learning” describes a method. “An administrative team cannot reconcile inconsistent maintenance records without substantial manual review” describes a potential need. A proposal still needs evidence that the problem exists and that resolving it matters. An interview record, documented workflow or appropriately shareable example may establish relevance without implying the interviewee has committed procurement funding.

Feasibility should explain what remains uncertain. If a prototype already exists, distinguish what it demonstrates from what the funded work would establish. A small demonstration on clean sample data may support the basic approach while leaving integration, representative data quality and reproducibility unresolved. State those boundaries openly. They give the reviewer a reason to fund the next experiment and make the resulting deliverable easier to assess. The UKDI fundable but not funded review explains why a positive assessment still differs from a funded project.

Define success before describing activity. “Run workshops and improve the algorithm” leaves the outcome uncertain. A stronger research plan identifies the question being tested, the evidence to be produced and the conditions under which the team would change direction. Avoid promising a universal performance figure from an experiment with a narrow scope.

Viability concerns the actual people, dependencies and resources needed. A founder named on several work packages is still one person. An external dataset that has not been licensed is a dependency, even if a demonstration used a similar dataset previously. Show where specialist effort enters the schedule and what happens if a dependency arrives late. Link cost to those activities so that the budget explains the plan. The FFI project cost analysis focuses on traceable assumptions before a supplier presents a delivery price.

The commercial case should continue beyond the project without claiming an assured purchase. Identify the likely next decision: a separate customer evaluation, an integration discussion, a commercial pilot or further research. Explain who would need to make that decision and what evidence this project would provide. A future revenue forecast becomes more credible when its prerequisites are visible.

An application can fail before the three criteria are fully assessed

UKDI's process includes an initial sift that checks matters such as scope, sufficient technical information, the required maturity, delivery timing and acceptance of terms. The guidance then describes expert assessment, moderation and a decision conference. This sequence means that a persuasive commercial story cannot compensate for a proposal that leaves the requested research undefined or exceeds the competition's limits. The application first needs to be assessable on the terms of that particular call. UKDI assessment stages

The writing implications are practical. Put the proposed work where the form asks for it. UKDI says the substantive desirable, feasible and viable answers must be in the text boxes, with image attachments limited to supporting purposes. An assessor should not have to reconstruct the proposal from a marketing deck or discover that an important technical explanation is only present in an attachment used to bypass a word limit. UKDI proposal submission instructions

This is also an argument for disciplined selection of evidence. A short example that directly supports the proposed method is more useful than a long list of loosely related projects. The company should explain the connection between earlier work and the new uncertainty: what has already been demonstrated, what changes in the proposed setting and why the team is equipped to investigate that change.

Build three connected arguments from one concrete research question

Consider a hypothetical company proposing research into the reconciliation of inconsistent maintenance records. Its desirability argument could begin with a documented administrative workflow: reviewers spend time resolving ambiguous identifiers before producing a service report. Evidence might describe where that burden occurs and why a more reliable process would matter. The value proposition should specify the beneficiary and the decision being improved, rather than start with the company's preferred algorithm.

Its feasibility argument would address the unresolved method. Perhaps the existing prototype performs well on consistently formatted records but has not been evaluated when several sources disagree. The proposal could define an authorised evaluation set and a method for comparing the new approach with a simpler baseline. The research question is whether the proposed method improves a defined part of the workflow under the stated conditions, not whether the company can demonstrate its existing interface.

Its viability argument would connect that evaluation to people, time and deliverables. Who prepares and reviews the input material? Who performs the comparison? When can the team interpret the result and produce a usable report? If a dataset or external reviewer is needed, the schedule should show the dependency and a credible alternative where appropriate. This makes the cost breakdown intelligible because the reader can connect expenditure to the evidence the project will produce.

The three sections should reinforce one another without repeating the same sales claim. Desirability explains why the problem matters. Feasibility explains why the proposed investigation can answer a useful question. Viability explains how the specific team will carry it out within the commission. A gap in one section often reveals a gap in the underlying project rather than merely a need for stronger language.

Explain what the result would change commercially

UKDI's assessor guidance asks about the route beyond the project, including further development or funding and suitable partners. That invites an explanation of the next decision, rather than a claim that the research itself completes commercial adoption. A company should identify which uncertainty the funded work removes and which organisation would need that evidence before taking the next step. UKDI exploitation and delivery criteria

In the administrative software example, a favourable result might justify a separately scoped integration evaluation with a customer or partner. A mixed result might show that the product is useful only for a narrower record type. An unfavourable result might prevent the business from committing to an expensive implementation model. Describing those possible consequences demonstrates that the project has a decision-making purpose and that the team understands what its evidence can establish.

The budget should reflect that purpose. If the central question can be answered through a bounded comparison, a proposal dominated by general product development may be difficult to justify. If a reliable answer requires substantial preparation of representative data, omitting that work makes the schedule less credible. The strongest proposal explains the resources necessary to obtain a useful result and shows how that result supports a realistic next commercial discussion. It gives assessors a research investment to evaluate, with its intended value and delivery logic visible throughout.

Before submitting, have a colleague read each section independently. They should be able to identify the beneficiary, unresolved technical question and deliverable without reconstructing them from marketing language elsewhere. This is an editorial method for preparing an assessable proposal, not an alternative scoring system. Meeting the criteria remains different from winning a competition with finite funding.