AlfaGov builds software for the senior officers accountable for government transformation, and keeps the record of how each decision was made — for whoever has to defend it next year.
Senior responsible owners move, delivery partners rotate faster, and the reasoning behind a programme is rarely written down in a form that survives either. The document lasts. Why an option was discounted, who agreed it, and what was overruled does not.
It shows up later as an inability to evidence anything. In April 2025 the Evaluation Task Force reported that two-thirds of the then Government Major Projects Portfolio, representing £456bn of whole-life cost, had no good-quality evaluation plan — with the weakest coverage in government transformation and ICT. The portfolio has since been reorganised and is smaller, so treat that as the last published portfolio-wide measure rather than a current one.
The same thing recurs across the published record: appraisal done to clear an approval rather than to choose between options, benefits defined once and never tracked, and risk registers maintained for assurance rather than for managing anything.
Source: Evaluation Task Force, evaluation of the Government Major Projects Portfolio, April 2025
Discrete products on one workspace. Each addresses a problem a senior responsible officer owns, and each is usable on its own.
Production
The case for the programme has to be made, and remade when it changes.
Structured exercises work through the strategic case — the case for change, the options, and why each was ruled in or out. The Green Book and the Five Case Model shape the structure. Every option, rejection and reason is kept on the record.
Covers the strategic case. It ingests existing financial models rather than generating the economic, commercial and financial cases.
In development
Benefits are defined at approval and nobody can evidence them afterwards.
Define benefits so they can be measured, carry them from the business case through delivery, and hold the evaluation plan the Treasury's own guidance asks for. The published evidence makes this the largest gap in the portfolio.
In development
Options are generated to justify a decision already taken, and it shows under scrutiny.
Generate and appraise a genuine long list, record the appraisal against stated criteria, and keep the reasoning for the options that were rejected — the part that is hardest to reconstruct later.
In development
The risk register is maintained for assurance rather than for managing anything.
Risk as a management tool with a reporting output, not a spreadsheet maintained for a review. Government's project data standard sets a monthly cadence for risk scoring, so the rhythm is becoming the expected one.
In development
Board papers are assembled by hand, every cycle, from sources that disagree.
Board and assurance reporting drawn from the programme's own record, so the paper reflects what the evidence says rather than what was available the night before.
NISTA's Government Reporting Integration Platform standardises what programmes report to the centre, and its assurance tooling supports reviews. That is reporting and assurance, and it is free to the organisations it covers. We do something different and narrower: we help build the case and hold the reasoning behind it, inside the programme, for the people accountable for it. If what you need is portfolio reporting to the centre, use the centre's platform.
Programme documents routinely hold options that have not been announced, commercial detail, and named officials. So the first question is not what the software does, it is where the material goes and who can reach it. Everything runs in one UK South Azure boundary, inference included. Sign-in is through your own Microsoft Entra directory, so access follows the conditional access policies you already enforce. There are no stored keys or passwords anywhere in the data path — access is by managed identity, which is a weaker thing to steal.
01
A working demonstration on your own problem takes days. Production software takes weeks. You see the thing before you commit to it.
02
The Green Book, the Five Case Model and government project delivery guidance shape the structure, so the output reads the way an assurance reviewer expects it to.
03
Every option, rejection and reason is kept. Not a log of what the software did, but a record of how the decision was reached.