Keep automation useful. Keep authority accountable.

Prove TDI is a Technology Decision Intelligence platform assisted by AI. Its governance model separates evidence from analysis, analysis from judgement, and recommendation from accountable approval. AI can help structure, compare, challenge and explain a decision, but AI output is not proof by itself and does not inherit decision authority merely because it is generated by the platform.

Prove TDI assists the decision process. It does not assume the legal, fiduciary, procurement or management authority of the organisation making the decision.

1. Decision governance

A defensible technology decision requires more than a ranked list. Prove TDI is designed to preserve who owns the decision, which requirements and constraints matter, what evidence supports the assessment, which conditions remain unresolved and who has authority to approve the resulting course of action.

• Decision ownership and approval authority remain attributable to accountable people or organisational roles.
• Requirements, constraints and material customer facts retain their decision context rather than being reduced to generic vendor criteria.
• Recommendations remain distinguishable from approval, procurement authority and final organisational commitment.
• Material assumptions, exceptions, contradictions and unresolved conditions should remain visible in the governed decision record.
• Where evidence changes materially, the prior decision state should remain distinguishable from the new assessment rather than being silently rewritten.
• Decision history and outcome review support learning from what was approved, why it was approved and what happened afterwards.

2. Evidence governance

Prove TDI treats evidence quality as a governed attribute. The platform is designed to retain provenance, source context, maturity and uncertainty so that a user can understand not only what a claim says, but what supports it and what remains unproven.

• Evidence retains source and provenance rather than being collapsed into an unexplained score.
• Evidence maturity uses the Prove TDI E0–E4 model to distinguish unsupported or low-maturity information from stronger forms of substantiation.
• Vendor-submitted material remains identifiable as vendor-controlled evidence until independently corroborated.
• Customer-provided facts, vendor claims, independent sources, expert input, inference and human judgement should remain distinguishable where that distinction is material.
• Missing, stale, contradicted or unsupported evidence remains visible when it could affect the decision.
• Evidence maturity and assurance are separate concepts and are not silently conflated.
• A newer source does not automatically erase the historical source or decision state it supersedes.
• Confidence should respond to the strength, recency, independence and consistency of the underlying evidence rather than to presentation quality.

3. AI governance and human oversight

AI is used as an analytical capability, not as an unaccountable decision-maker. Its role is to improve the speed, breadth and consistency of evidence work while keeping material facts, uncertainty and accountable authority visible to the people responsible for the decision.

• AI may assist with structuring, classification, synthesis, comparison, challenge, gap identification and explanation.
• AI-generated analysis is not automatically promoted to established evidence.
• Material claims should remain grounded in attributable evidence or be represented as inference, uncertainty or an evidence gap.
• Human review is retained where context, trade-offs, exceptions, conflicts or accountable approval are material.
• Model confidence is not treated as evidence confidence.
• Changes to models, prompts, workflows or providers should not silently change historical decision authority or the provenance of established evidence.
• Automated processing does not remove the customer's responsibility to determine whether a recommendation is appropriate for its legal, procurement, operational and organisational context.

AI transparency and regulatory boundaries

Prove TDI does not use a blanket statement such as “AI Act compliant” to imply that every customer use case, deployment role or future AI capability has the same regulatory status. AI governance and transparency requirements depend on the role Prove TDI performs, the capability being used and the customer's intended context. Those obligations should be assessed for the relevant use case and production scope.

Prove TDI is designed to support role-specific AI governance and transparency obligations, with human decision authority retained by the customer where accountable authority is required.

4. Commercial independence

Commercial participation must remain separate from evidence quality and decision outcome. This separation is central to Prove TDI's role as an independent Technology Decision Intelligence platform.

• Vendors cannot buy a higher evidence maturity level, confidence score, ranking or recommendation.
• Paid verification or assurance activity cannot convert unsupported information into stronger evidence without the required substantiation.
• Vendor, partner, expert or referral relationships do not determine the recommended option.
• Vendor-supplied information retains attribution and does not become independent evidence merely because it appears on Prove TDI.
• Claiming a vendor profile can establish authorised participation or affiliation; it does not make every vendor claim independently verified.
• Expert conflicts and relevant commercial relationships should be disclosed or governed where they could affect independence.
• Sponsored activity, where offered, must remain outside the governed calculation of evidence maturity, ranking, recommendation and approval.

Privacy and data-governance principles

Prove TDI treats privacy as part of governance rather than as a separate marketing claim. Personal-data handling should be proportionate to purpose and assessed alongside the public Privacy Notice, customer scope and applicable contractual requirements.

• Lawfulness and transparency should be established for relevant personal-data processing.
• Data should be collected and used for defined purposes rather than retained because it may become useful later.
• Data minimisation should limit personal information to what is necessary for the relevant purpose.
• Accuracy, retention and deletion requirements should be handled according to the relevant data and customer context.
• Integrity, confidentiality and access control remain security as well as privacy concerns.
• Accountability requires Prove TDI to distinguish demonstrated controls from assumptions, roadmap items and unsupported compliance claims.

Governed boundaries

The following boundaries are deliberately explicit because they prevent automation, commercial influence or presentation quality from being mistaken for proof.

• Evidence is not the same as a claim. A statement retains its source, maturity and assurance context.
• Analysis is not the same as evidence. AI or human synthesis may interpret evidence without becoming the underlying proof.
• Confidence is not certainty. Material gaps and contradictions remain part of the decision.
• Recommendation is not approval. Accountable authority remains with the relevant customer decision-maker.
• Verification is not sponsorship. Commercial participation cannot purchase a decision outcome.
• Readiness is not certification. Alignment, preparation or roadmap activity is not represented as independent attestation.

Governance for enterprise and public-sector review

Organisations may require additional governance evidence for procurement, regulated use, public-sector accountability, internal audit, risk management or AI oversight. Prove TDI supports scope-specific review so that applicable controls, decision roles, data conditions, assurance requirements and unresolved gaps can be established before deployment rather than inferred from generic website language.