Assessment intelligence
Matrix
Build the framework, run the assessment, defend the score. Matrix carries a maturity model from its first statement to a certificate an awarding body can stand behind.
Counted in the running platform, September 2026
The problem
Most maturity assessments are a consultant's spreadsheet with a logo on it.
The model lives in one workbook, the evidence in an inbox, the score in a slide — and nobody can reproduce last year's number. Matrix makes the framework a first-class object: versioned, cloneable, exportable, and every score traceable to the statement and the evidence behind it.
Build the framework
Design perspectives, dimensions, criteria and statements — or let the AI builder draft them from a standards library and edit from there. Clone a framework to version it; export it to hand it over.
Run the assessment
Multiple assessors, a consensus view to reconcile them, evidence attached to the statement it answers, plus tokenised surveys and interviews that need no login.
Write the report
A maturity report with current-state practices, gap analysis and recommendations written per perspective — editable in place, regenerable, and exported as a bilingual PDF or a slide deck.
Award and revoke
A full accreditation lifecycle: applications, review, five certificate levels, expiry, renewal and revocation — ending in a QR-verifiable certificate and a public registry.
What it assesses
Frameworks are data, not code.
Every framework on the platform was built in the product rather than shipped as a code seed — which is why a school quality model and a governance-to-value model can run side by side without either being a fork.
Maturity models
Compliance frameworks
Reference standards built in
Maturity scale
One framework, in numbers —Comprehensive Maturity Assessment
Assessment to accreditation
Most tools stop at a score. Matrix issues the certificate.
Built for awarding bodies
Applications, review, award, expiry, revocation — and a public register anyone can check.
Award rules are machine-checked against the framework: a minimum overall score, a minimum on a named dimension, mandatory evidence, a completed workflow. The certificate carries a QR code that resolves to a public verification page, so a claim on a letterhead can be checked in a second.
The AI question
Pluggable, per tenant — including a model you host yourself.
Matrix runs on Anthropic Claude, OpenAI or a self-hosted Qwen deployment, selected per organisation without a code change and with provider keys encrypted at rest. For a buyer who cannot let assessment data leave their network, the self-hosted option is the difference between a project and a policy problem.
What the AI actually does
It drafts frameworks and writes the report — in Arabic as readily as English.
Generating a framework from a standard; writing the current-state, gap and recommendation narrative for each perspective; drafting bilingual survey and interview questions; transcribing recorded interview answers; and an in-product assistant your organisation names and brands itself.
Arabic is in the data model
Every statement exists in both languages. So does every report.
Not a translated menu bar: the framework names, the 5,803 statements, the survey questions and the AI-written analysis all carry an Arabic column, and the PDF renders right-to-left with proper Arabic typesetting. The deliverable is Arabic, not just the interface.
Arabic covers effectively all of the production interface. A small number of the newest administration screens still show English labels inside the Arabic UI; those are being closed.
Let’s engineer what comes next.
Bring us the framework you already use, or the one you have been meaning to write down. We will show you both running.