AI Control Plane · Evidence
What does one decision record prove?
Follow a governed action, inspect its record and check what the evidence can establish. Start with the public ACP examples below; each includes its source, date and limits.
Product access opens the ACP sign-in page. The public examples need no account.
Looking for the earlier security-training research? Read the historical methodology or its public references
Public ACP examples
Inspect the action, decision and record.
These published examples let you inspect ACP evidence before evaluating it for your own workflow. Read each boundary alongside the record: the checks below answer different questions.
They cover the demonstrated paths through ACP. Activity bypassing ACP remains outside its control path. These examples provide technical evidence for review within that boundary.
- Inspect a governed decision record →
Six chained records from one governed change, with the exported JSON and a recipe to recompute the chain.
Captured 7 August 2026 · Public ACP example
Boundary: Checks internal consistency. An unkeyed hash chain does not prove origin or prevent someone from rewriting the whole bundle.
- Signed evidence pack (Ed25519) →
Download the demonstration bundle, public key and verifier. Compare the key fingerprint with the published trust anchor.
Captured 7 August 2026 · Demonstration key
Boundary: Verifies against the supplied demonstration key. It does not establish legal identity or custody of a production key.
- Ninety-second demonstration →
Watch an allowed action, a refusal outside a delegated mandate and a decision-chain check in a synthetic scenario.
Produced 8 August 2026 · Synthetic demonstration
Boundary: The organisation is fictional. This is a demonstration of the shown workflow, not a customer outcome or proof of coverage outside ACP.
IntelXview Limited has been selected to participate in Cohort 2 of the FCA AI Supercharged Sandbox. Participation does not imply FCA endorsement, authorisation, registration, approval, or validation of IntelXview Limited or its products or services.
Historical context · LimitedView
Security-training methodology and references.
Earlier material discussed incident-based training, assessment before and after learning, and delayed retention checks. This archive preserves that methodological context separately from ACP product evidence.
Historical outcome figures, cohort sizes and comparisons have been omitted pending verification of their underlying sources. The public papers below provide background; they do not substantiate internal LimitedView results or ACP outcomes.
To assess a training result, ask for the source dataset, sampling and deduplication rules, assessment design, timing and uncertainty. Course completion, self-reported confidence and measured retention describe different things.
Historical bibliography
Sources to read in their own context.
These learning-science papers and industry reports are retained for reference. Each source has its own population, methods and scope.
- Murre, J. M. J., & Dros, J. (2015). Replication and analysis of Ebbinghaus' forgetting curve. PLOS ONE, 10(7), e0120644. Read the paper
- Proofpoint, State of the Phish 2024; SANS, Security Awareness Report 2024. Proofpoint report library · SANS report library
- Mather, M., & Sutherland, M. R. (2011). Arousal-biased competition in perception and memory. Perspectives on Psychological Science, 6(2), 114–133. Read the paper
- Roediger, H. L., & Karpicke, J. D. (2006). The power of testing memory: Basic research and implications for educational practice. Perspectives on Psychological Science, 1(3), 181–210. Read the paper
Related material: ACP workflow examples · evidence-pack overview · ask about a source.
Which workflow do you want to evaluate?
Explore ACP, or tell us which action and evidence matter to you. Any paid PoC needs a separate written scope and price that you choose to accept.