NexArt adds independent timestamping and confidential mode to AI execution records

NexArt has added external DigiCert timestamping and a Confidential Execution Mode to its Certified Execution Records, aiming to let organisations prove what AI systems executed without exposing sensitive data. The update is designed for audits, disputes and regulated workflows where verifiable evidence matters.

Why it matters: - NexArt’s update strengthens AI execution records for audits, disputes, incident reviews and regulated workflows. - The new features are designed to improve evidentiary value while reducing the need to expose sensitive, proprietary or regulated data. - Organisations can verify records without requiring a NexArt account or trust in the original operator.

What happened: - NexArt announced two enhancements to its Certified Execution Records on July 7, 2026, in London. - The company added independent RFC 3161 timestamping through DigiCert’s public timestamp authority. - NexArt also introduced Confidential Execution Mode for organisations that need verifiable records without exposing sensitive information.

The details: - NexArt describes itself as an AI evidence layer that produces tamper-evident, cryptographically verifiable records of what AI systems and agents actually executed. - Certified Execution Records are intended to help teams prove what happened to auditors, clients, regulators, partners or internal reviewers. - Any third party can independently verify a Certified Execution Record using standard cryptographic methods. - The new timestamping feature adds an external trusted time source to the record. - NexArt says that strengthens the record’s value in audits, disputes, incident reviews and regulated workflows. - Confidential Execution Mode uses keyed HMAC-SHA256 commitments. - The mode allows records to be created and verified without the verification node seeing plaintext inputs or outputs. - Authorised parties can verify the integrity of the execution record while keeping sensitive data private. - Jeremy Bouedo, NexArt founder, said the company released its first SDK six months ago after a year of engine development. - Bouedo said the new additions continue the company’s focus on building evidence that stands on its own. - Bouedo said NexArt is committed to creating records that do not require trust in the producer or in NexArt itself. - NexArt records and verifies what was executed by an AI system. - NexArt does not provide governance, model evaluation, output scoring, fairness assessment or legal compliance certification. - Those responsibilities remain with the organisation deploying the AI system. - Certified Execution Records can be verified at verify.nexart.io.

Between the lines: - NexArt is positioning its product as proof infrastructure, not a broader AI governance platform. - The confidential mode suggests the company is targeting regulated industries and proprietary workflows where transparency and secrecy must coexist. - The emphasis on external timestamping points to a push for records that can stand up in higher-stakes reviews without relying on internal trust alone.

What's next: - Organisations using AI systems can now test the updated verification flow through NexArt’s public verifier. - NexArt is likely to lean on the new timestamping and confidential mode as it expands use cases in regulated and sensitive environments. - The company’s next challenge is adoption: proving that independent verification is useful enough for auditors, legal teams and operators to build around it.

The bottom line: - NexArt is trying to make AI execution records more trustworthy to outsiders while keeping sensitive data private.