A plain guide to the grant applicant risk score for trustees, auditors and panel members. Written for charities, community funders, research organisations and central government.
The grant applicant risk score in Flexigrant’s AI Due Diligence Checker is a percentage produced by a fixed formula, not by AI. It scores five factors from public register data: filing compliance, financial health, insolvency risk, charges and debt, and governance. Low is 0 to 30%, Medium 31 to 60%, High 61 to 100%. The score is advisory.
If you sit on a grants panel or audit one, you will be shown this number. This article explains what it is built from and why a score can be out of 80. It covers what the three bands mean and what a red flag is. It also sets out what the score cannot do.
Because the question has changed. It is no longer “did you check?” but “how did you check, and was it the same for everyone?” Boards, auditors and the public expect a consistent, written answer for every applicant. A score from a fixed formula gives you one. Every applicant is measured against the same five factors, at the same depth. The result sits with the application, in the system that runs your whole grant process.
That is the value for the person who has to defend a decision later. The score does not make the decision. It shows, in one place, that the same check was applied and what it found. The breakdown and the red flags show how the number was reached, and every register field links to its source record.
No. A fixed formula calculates the score. The same inputs always give the same result. The rules do not change between applicants, and they only change over time when Flexigrant releases a change. This matters for audit. You can explain a formula to a trustee. You cannot explain why a language model felt a certain way about an applicant.
AI does appear in the checker, in one clearly marked place. Azure OpenAI writes a plain-English summary and analysis notes to sit alongside the score. Those are labelled as AI-generated, and you are asked to check them against the source records. The number itself never passes through the AI.
The formula scores five factors drawn from Companies House, the Charity Commission for England and Wales and OSCR.
Filing compliance covers whether the organisation is keeping up with its statutory filings. Financial health covers the financial data the registers hold. Insolvency risk covers insolvency data, such as administration or liquidation events. Charges and debt covers charges, which are legal claims a lender holds over assets as security for a debt. Governance covers the directors or trustees on record.
Each factor scores separately, and you see the breakdown, not just the total. If one factor is pulling the score up, the breakdown shows you which. That is where a panel conversation should start.
Not every register holds every kind of data for every organisation. An unincorporated charity, for example, has no Companies House record. When the checker has no data for a factor, it excludes that factor rather than scoring it as zero or as perfect. If one factor is excluded, the score is calculated out of 80 instead of 100.
The percentage you see is the points scored against the points available. So a score can be labelled out of 80 and still show as a percentage within the same three bands. The screen tells you which factors were assessed. For an auditor, that is the first thing to note. A score with three factors assessed rests on less evidence than one with five.
The bands are fixed. Low is 0 to 30%, Medium is 31 to 60%, and High is 61 to 100%. A higher percentage means more risk indicators in the register data, not a better result.
A Low score tells you the registers show few concerns. A Medium score tells you to read the breakdown before deciding. A High score tells you the register data raised several concerns, and your policy should say what happens next. None of the bands is a decision. Your funding policy decides what each band requires, whether that is a second review, a request for more information, or a decline.
A red flag is a specific finding in the register data that the checker calls out on its own. It is shown separately so it is not lost inside the overall number. The score summarises. A red flag points. It tells you where to look in the source record and what to ask the applicant about.
Red flags are shown with the breakdown, and each links back to the register that raised it. Treat them as questions, not verdicts. A flag may have a good explanation, and the applicant is often the right person to give it. Your process should record the flag, the explanation, and the decision you took.
Read it in this order. First, note the band and which factors were assessed. Second, open the breakdown and find the factor carrying the most weight in the result. Third, read the red flags and follow each link to its source record. Fourth, read the AI-generated summary as a guide to the record, then check anything it says against the register field it describes. Fifth, apply your policy and record the decision.
That sequence gives you a documented check that another person can repeat. It also keeps the responsibility where it belongs. The checker is advisory. It supports your due diligence process and does not replace it.
You run the check from the Manage Grants menu in Flexigrant, from Fluent Technology. You enter the organisation name and, ideally, its company or charity number, because registered names often differ from working names. One search retrieves the records from Companies House, the Charity Commission and, when you enter a Scottish charity number, OSCR. The score, the breakdown, the red flags and the labelled AI summary appear together on one screen. The checker is built into Flexigrant and included in every subscription. As of 2026, more than £6.5 billion has been awarded through the platform.
Read the full AI Due Diligence Checker feature page
Book a demo, and we will run the checker on organisations you name and walk through the breakdown with you.