Summit Cognitive Launches the Decision Receipt, Making AI-Assisted Decisions Defensible and Independently Verifiable
A governed decision platform — live in production, with a public key and ledger anyone can use to verify the record.
ALMA, Colorado — June 26, 2026 —
Summit Cognitive, Inc. today announced that its governed decision platform is live in production, issuing the Decision Receipt: a cryptographically signed, replayable, independently verifiable record of why an AI-assisted action was allowed, revised, blocked, or escalated. The platform is operating now at decrec.summitcognitive.ai, where a public key and a ledger are published so that any third party can verify a receipt without trusting the company that produced it.
Most organizations can tell whether an AI system produced an output. Almost none can prove that the output was decision-valid — authorized, evidence-linked, and policy-compliant — captured at the moment the decision was made. Summit Cognitive closes that gap. The Decision Receipt turns an AI-assisted decision into a governed artifact that shows its work, holds up under audit later, and can be contested by an affected party.
The platform operates at runtime: every governed action is evaluated and signed before the output leaves the system, rather than reviewed on a later cycle. Signatures use Ed25519, and the policy layer denies by default — an action is allowed only when it clears the controls in place. Because the record can be replayed against what was actually known at the time, a decision can be reconstructed and reviewed instead of merely narrated.
The numbers below are public-level, audit-time figures and are independently checkable against the live ledger:
Signed receipts in production
3,000+
Receipt acceptance (deny by default)
98.6%
Replay determinism
99.9%
Governed agents in production
9
Signatures
Ed25519, public-key verifiable
Codebase
52 canonical packages · 5 services
Engineering metrics are audit-time figures captured in late May to early June 2026; some have since advanced. The product is built, deployed, and in active pilot motion — not a model wrapper, not a UI overlay, not a research project.
Summit Cognitive describes the layer it operates in as runtime admissibility for AI in high-stakes decisions — the record that AI governance, model safety, and audit all need underneath them. The company's broader argument is set out in the forthcoming book Admissible Reality by founder Brian C. Long, and in the company's Dispatches essay series and the Warrant podcast.
On the federal side, the company's credentialing is in place: a Unique Entity ID (UEI) has been assigned, the company is registered on SAM.gov, and a CAGE code has been assigned. Summit Cognitive's provisional patent application, No. 64/034,952, was filed on April 10, 2026; the non-provisional conversion is due by April 10, 2027. The patent is pending, not granted.
The Decision Receipt is available to verify now. Affected parties, auditors, and prospective design partners can examine the public key and ledger directly at decrec.summitcognitive.ai.
About independent verification
Verification does not require an account, a credential, or Summit Cognitive's cooperation. The public key is published; the ledger is published; a receipt either verifies against them or it does not. That is the point: the company is not asking to be believed. [[ TK — confirm with Brian: link to step-by-step verification instructions on the press/verify page ]]
— ENDS —