Compliant AI — Established 1994
Audit-proof AI for credentialing bodies.
Four fixed-scope engagements that produce the evidence your audit cycle expects. Principal-led delivery on managed AWS foundations with Claude on Bedrock. Governance built in from the first commit. Scoped to your programme in one conversation.
Trusted by global enterprises including Siemens, Unilever, Microsoft, IBM, Hitachi, and the Society of Actuaries.
Four engagements. Each producing audit-ready evidence.
Fixed scope, Principal-led delivery, 90-day rhythm. Each one maps to a named regulatory or professional-standards obligation. Each one can be scoped to your programme in one conversation.
Governance Pack
The audit-ready evidence base under the EU AI Act, ISO 42001, and the AERA–APA–NCME Testing Standards. AI register, risk classification by stakes, audit pack components, and a three-tier governance operating model — delivered alongside the platform you already run.
Read the deep dive →Grader Analysis & Calibration
Inter-rater agreement evidence across your examiner panel, drift detection with trigger thresholds, subgroup bias monitoring, and pre-launch calibration with synthetic submissions. Reliability evidence your regulator, accreditor, and standards body will accept.
Read the deep dive →Multi-modal Feedback at Scale
Every learner receives personalised PDF, audio podcast, avatar video, and competency heat-map feedback — produced automatically from the grading decisions your graders already make. One grading process, four channels of feedback, every candidate.
Read the deep dive →The Precision Report
Quarterly governance evidence produced on the audit cycle, not assembled retrospectively when the audit arrives. Reliability, fairness, drift, bias, override logs, incident summary — against your assessment data, mapped to the standards you must meet.
Read the deep dive →How we deliver
Principal-led. Anthropic-first on AWS. Governance from the first commit.
Principal-led delivery
You meet the senior consultants in the sales conversation and you work with the same senior people in delivery. No staff substitution after contract. Thirty-plus years of regulated-industry experience on every engagement.
Your data, your perimeter
Claude on Amazon Bedrock keeps your prompts, context, and governance evidence inside your AWS environment. The difference between an architecture that survives an audit and one that does not.
90-day rhythm
Every engagement defines the workflow in weeks 1–3, builds a working artefact with governance wired in during weeks 4–8, and measures against the metric in weeks 9–12. If the first delivery cannot ship in 90 days, the scope is wrong.
Thirty years of regulated-industry delivery
Trusted by global enterprises including Siemens, Unilever, Microsoft, IBM, Hitachi, the Society of Actuaries, and credentialing bodies across nineteen-plus countries. Twenty-five per cent of revenue reinvested in R&D; forty-seven-plus patents and innovations in the engineering underneath.
Before the call
Questions we hear most often.
No. The four engagements are designed to sit alongside whatever you already run — Questionmark, Surpass, Cirrus, or your own build. The evidence artefacts, the governance operating model, and the reliability work are platform-agnostic.
The EU AI Act (high-risk systems under Annex III), ISO 42001 and 23894, the AERA–APA–NCME Standards for Educational and Psychological Testing, the NCCA Standards for certification programmes, and UK/EU GDPR where candidate data is in scope. The EU AI Act piece sets out the detail →
General-purpose AI governance platforms do not speak credentialing. Construct validity, inter-rater agreement, accreditation evidence, and the Testing Standards are absent from their vocabulary. Our work is sector-specific by design — the same engineering discipline, applied to the programmes your reputation depends on.
Every engagement is led by a named Principal Consultant with thirty-plus years of regulated-industry experience. You meet them in the sales conversation and you work with the same person in delivery. No junior substitution after contract. No eighteen-month discovery phases.
Fixed scope, fixed fee, ninety-day rhythm. Scope is agreed in week one against a defined deliverable. No open-ended consultancy bills. Pricing is scoped against your programme size and portfolio — not against a published rate card that does not fit your specific circumstances.
Book a thirty-minute call with a Principal. Cameras on, no slides. We will map your current position against the four engagements and tell you honestly which fits, which is premature, and which you do not need.
Talk to a Principal.
Thirty minutes. A named Principal Consultant on the call. We will map your current position against the four engagements and tell you honestly which fits — including declining if there is not a fit. Most engagements move from first call to signed statement of work in three to five weeks. Cameras on. No slides. Plain answers.