Top tier of the Reign subscription. Adaptive recalibration, drift detection at scale, residual risk and trust scoring. Continuous adaptive assurance.
Reign Continuous is the top tier of Continuous Operational Assurance for Enterprise AI. Both assurance questions are now answered continuously, before execution and after execution, across the whole AI portfolio, with the operating model that lets the customer sign against the answer on any given day. Outcome telemetry feeds back into trust scoring. Drift triggers recalibration. The audit-attestation flywheel runs continuously, the Palantir model for AI governance in regulated industries.
“How do we operate adaptive assurance across a scaled portfolio of agents and workflows?”
The runtime control plane and the written agreement that spells out the rules each agent must operate under are both in place. Reign Continuous extends them with adaptive recalibration, portfolio-scale trust scoring, and a regulator-ready operating model.
Policy is not a write-once artifact at this tier. Outcome telemetry from the audit ledger feeds back into the policy layer, drift signals from the validation harness trigger recalibration of the controls, and the policy set evolves on the cadence named in the written agreement that spells out the rules each agent must operate under. Recalibration is logged the same way every other gateway decision is logged. The policy version that was in force when an action ran is preserved in the ledger record for that action. Forensic reconstruction stays intact across recalibrations. The flywheel runs continuously and the evidence corpus stays defensible.
Drift on inputs, drift on outputs, drift on embeddings, drift on tool-call distributions, drift on the residual risk score itself. At Continuous, drift is monitored continuously across the whole portfolio, not workflow by workflow, and material drift triggers the recalibration loop above the moment it is detected, not on a scheduled review. This is the surface that lets a customer at Level 5 maturity hold a defensible posture across many workflows, many business units, many regulators, simultaneously.
Each model, each agent, each workflow carries a residual risk score and a trust score that is continuously updated from outcome telemetry. The board sees the portfolio rolled up. The risk and audit teams see it by workflow. The operator sees it per call. Trust scoring is by construction. The score is a function of the evidence in the ledger, the validation history, the drift signal, and the outcome telemetry. It is not a separate model the customer has to defend. The defensibility comes from the underlying corpus.
The whole operating posture is set up to answer regulator questions in the timeframes regulators set, not in the timeframes a reconstruction effort would need. SR 26-2 examiner asks for documented validation of a model that produced a credit decision in March, the answer is in minutes, with replayable evidence, against the policy version that was in force at the time of the decision. This is the audit-attestation flywheel running continuously. The Palantir model for AI governance in regulated industries, forward-deployed operating posture, attestor-neutral substrate, deep customer integration, defensible against any regulator the customer is in scope for.
Outcome telemetry feeds back into trust scoring. Drift triggers recalibration. Recalibration is logged the same way every other gateway decision is logged. The flywheel runs continuously, and the evidence corpus stays defensible.
Captured at the call layer, committed to the ledger.
Updated continuously per model, per agent, per workflow.
Material drift across the portfolio triggers the loop.
Controls evolve on the cadence named in the written agreement that spells out the rules each agent must operate under.
The policy in force at every decision is preserved in the ledger record.
Schematic. The diagram is abstract by design. The internal architecture is documented in the assurance contract, not on this page.
The three subscription tiers are maturity stages, not packaging options. Continuous sits at the top of the model, where the operating model itself adapts in response to evidence.
Both assurance questions are answered across multiple workflows. Evidence packs are on a contract cadence. The audit committee has a board-grade view. This is where the second tier ends.
The operating model itself adapts in response to evidence in a way that stays defensible to regulators and auditors. Level 5 is not a feature, it is an operating posture.
Level 4 is scaled autonomy, the level Assurance takes the customer to. Level 5 is continuous adaptive assurance, the top of the maturity model, where the operating model itself adapts in response to evidence in a way that stays defensible to regulators and auditors.
Customers at Level 5 are usually running AI across many business lines, in many jurisdictions, under multiple framework regimes at once. Reign Continuous is the tier that holds the posture together at that scale, delivered as services-as-software.
Read the Autonomy Assurance Maturity ModelThe engagement funnel has four stages. Executive Assurance Briefing, Runtime Risk and Governance Assessment, Focused Pilot, and Platform Rollout. Continuous is the highest tier of the Reign subscription that customers step up to after Platform and Assurance, or enter directly when the portfolio is already at scale. The annual subscription includes a full iTmethods operating partner team, a named assurance lead, a recalibration lead who owns the closed-loop policy practice, and an executive sponsor from iTmethods at the audit and risk committee cadence.
iTmethods supports the customer in running Reign across the whole AI portfolio. Embedded, not advisory.
Owns the written agreement that spells out the rules each agent must operate under, end to end, plus the pack delivery cadence and the external attestor relationship.
Owns the policy recalibration loop, the drift signal triage, the portfolio-scale trust scoring practice, and the operating-model documentation the regulator reads.
From iTmethods, at the audit and risk committee cadence. On demand with the regulator and the external attestor.
The path into Reign Continuous is typically a stepped progression through Platform and Assurance. Customers occasionally enter at Continuous when the AI portfolio is already at scale and the regulator timeline does not allow a slower ramp. Either path starts the same way. Briefing, Assessment, Pilot, then Platform Rollout at the Continuous tier.
The first conversation is the Executive Assurance Briefing. We map the AI portfolio, the jurisdictions, the framework set, the recalibration cadence the operating model needs to support, the executive sponsorship at the customer, and the external attestor relationship. The Continuous tier subscription is sized to the portfolio, not to a workflow count.
Reign does not replace enterprise governance, legal review, model risk management, or audit. It gives those functions runtime evidence, exception visibility, and continuous assurance signals they can act on.