Product portfolio

Four systems. Three enterprise outcomes.

CAI is organised around Decide, Assure and Transform — consequential work where system truth, deterministic logic, AI reasoning and human authority must remain traceable.

Explore the portfolio Discuss deployment

Current systems

Bounded products for work where evidence and authority matter.

01DECIDE

DecisionReady

Prepare material enterprise decisions before leadership commits.

Structure the ask, evidence, assumptions, blockers, dependencies and leadership questions into a governed decision-readiness process.

Operating question

Is this decision sufficiently evidenced and challenged to progress?

Core output

Readiness, blockers, missing information, contradictions, risks, challenge questions, pre-review actions, human review and decision lineage.

Human authority

The accountable decision owner retains approval, rejection, escalation and override authority.

Product boundary

DecisionReady prepares and challenges the decision context; it does not make the underlying business decision.

Open DecisionReady Product details

02ASSURE

AuditEvidence

Make the evidence-to-professional-decision chain explicit, current and defensible.

Govern evidence request, exact evidence use, quality, sufficiency, exceptions, professional conclusion and sign-off without collapsing AI output into professional authority.

Operating question

Can another reviewer understand and reperform how this professional decision came into existence?

Core output

Versioned evidence context, reliance, evidence-quality review, sufficiency, exceptions, human conclusion, sign-off and currentness.

Human authority

Auditors and reviewers retain reliance, review, exception, conclusion and sign-off responsibility.

Product boundary

AuditEvidence is not an autonomous auditor or a replacement for the entire GRC and repository stack.

Open AuditEvidence Product details

03TRANSFORM

PQM

Run post-quantum migration as a governed operating programme.

Connect discovery, crypto inventory, dependencies, prioritisation, migration planning and verification in an evidence-led migration operating system.

Operating question

What must change, in what order, on what evidence, and how will remediation be verified?

Core output

Workspace posture, findings, dependencies, priorities, migration plans, approvals, verification evidence and governed state.

Human authority

Consequential migration approvals remain explicit human decisions; approval does not imply execution.

Product boundary

The product does not use synthetic readiness scores or imply unsupported production scanning, remediation or certification.

Open PQM Product details

04TRANSFORM

Controlled Conversion

Move critical spreadsheet logic into a controlled application without losing the logic that produced the number.

Specify the rules, rebuild them with controls, replay historical periods, reconcile every difference and transfer trust only when the evidence supports it.

Operating question

Can this spreadsheet-dependent calculation be converted without losing its intended logic or control basis?

Core output

Controlled application, signed specification, historical replay report, explained differences and operating evidence pack.

Human authority

The accountable process owner signs the specification and reconciled differences; material overrides remain maker-checker.

Product boundary

Current offer is calculation and reconciliation focused, with file-based ingestion and client-owned or client-approved hosting.

Product details

What the products share

Govern the basis before you automate the consequence.

System truth: preserve the authoritative facts, state and source evidence.

Deterministic rules: policy, thresholds and hard constraints stay explicit.

Bounded AI: model reasoning assists interpretation without silently becoming authority.

Human authority: consequential review, approval, reliance and sign-off remain attributable.

Provenance: the basis, changes and resulting governed state remain inspectable.

Operating boundary

CAI products support governed enterprise work; they do not erase accountable human authority.

Deterministic truth, model assistance, uncertainty and human decisions remain visibly separate so the operating basis can be inspected rather than merely trusted.

See the trust model

Discuss a fit

Have consequential work that resembles one of these operating patterns?

Start with the work itself: authoritative evidence, deterministic constraints, judgement points, decision rights and the state that must remain traceable.

Discuss deployment