INDICIO Proven AI
The trust layer for AI.
Agents can’t prove who they are. Users can’t tell which agents are legitimate. Nothing governs what happens when agents talk to each other. Indicio Proven® AI solves all three, giving AI agents verifiable identity on the same credential infrastructure already deployed at national-scale borders and global banks, and providing the capacities needed to make agentic identity and the agentic web possible.
AI without identity is AI without trust.
Agents handle loan applications, bookings, and account changes on OTPs and security questions that deepfakes and social engineering defeat. No cryptographic proof the user is real.
Nothing lets a user confirm the agent they’re talking to is legitimate. Phishing and impersonation are trivial to run and impossible to spot without verifying the agent’s identity.
When agents delegate tasks, share data, and execute transactions with each other, nothing verifies authority, scopes permissions, or keeps an auditable trail.
Mutual authentication for humans and machines.
The same credential technology that clears travelers through borders and verifies bank customers, applied to every AI agent and every interaction.
At the start of an interaction, the user consents and taps to share a verifiable credential from their wallet.
The agent verifies the user’s credential is authentic. At the same time, the user’s software verifies the agent is legitimate. Proven both ways before any data moves.
Structured, pre-verified data moves straight to the agent: name, account, permissions, context. No manual entry, full consent, full audit trail.
“AI is moving into personal, high-stakes spaces where identity, consent, and data control are essential. Proven AI gives developers the tools to build agents that can prove who they are and prove who you are, with your permission.”
Ken Ebert, CTO, Indicio
New whitepaper
Identity, Commerce, and the Agentic Web
AI agents are starting to act, buy, and share data on people’s behalf, across systems that have never met them. This paper lays out how verifiable credentials give the agentic web the trust layer it needs, proving identity, consent, and authority before any data moves.
- Why API keys and access tokens cannot secure agent-to-agent interaction
- How agents that have never met establish trust on first contact
- The standards and regulators converging on verifiable credentials
Proven AI for travel and KYC.
Two products shipping today: agents that authenticate travelers with Digital Travel Credentials, and agents that verify banking customers with reusable KYC.
The first product that lets AI agents verify digital passports. Travelers present a Digital Travel Credential alongside loyalty and other credentials, and structured, pre-verified data flows straight to the agent. No manual entry of names, flight details, or PNRs, which cuts input errors, failed authentications, and support escalations.
Bank and fintech agents verify customer identity from a credential in the customer’s wallet, with no passwords, document review, or perception-based biometrics. Customers verify once and reuse the credential across institutions. Because proof is cryptographic, not visual, it’s a structural defense against deepfakes and synthetic fraud, with auditable records for KYC, AML, and CIP.
The missing layer: governance for agent networks.
Indicio DEGov publishes machine-readable governance files that tell every participant in a network, issuers, holders, verifiers, and agents, who is trusted, what permissions exist, and which interactions are authorized. This is the capability Gartner recognized as unique to Indicio: trust across autonomous systems, enforced cryptographically.
Governance files are published by the network’s natural authority. Every agent knows who to trust before an interaction begins.
Authority has levels. An agent can delegate to another agent only within its scoped permissions.
Agents from different organizations interact in a shared trust framework. No bilateral integrations, no central broker.
Every presentation, permission check, and delegated action is logged, timestamped, and cryptographically verifiable.
Why the analysts are paying attention.
Market Guide for Decentralized Identity, 2025.
Building the trust infrastructure for agentic AI at scale.
The venture arm of NEC Corporation, backing decentralized identity for AI agents.
On the ProvenAI launch and the NVIDIA partnership.
Questions & answers
Proven AI, explained
What is Proven AI?
Proven AI is a trust layer for AI agents. It gives an agent a verifiable identity and verifiable proof of what it is authorized to do, so the systems and people it interacts with can confirm who is acting and on whose behalf, rather than taking an unverified claim at face value.
How does an AI agent prove its identity?
The agent holds verifiable credentials issued by a party that stands behind it, such as the organization that deployed it. When the agent acts, it can present those credentials, and the other side can check them cryptographically against the issuer's published keys. That turns "trust me, I'm an authorized agent" into something a counterparty can independently verify.
How do you control what an agent is allowed to do?
Authority can be expressed as a verifiable credential that scopes what the agent may do and on whose behalf, and that can be checked at the point of action. Because the authorization is a credential, it can be issued, presented, and revoked, which gives you a way to grant and withdraw an agent's authority rather than relying on standing access.
What problem does verifiable identity for agents solve?
As AI agents begin to transact and act for people and organizations, the open question is accountability: which agent did this, who was behind it, and were they permitted to act. Without verifiable identity, an agent's claims about itself cannot be checked, which makes impersonation and unauthorized action harder to detect. Verifiable credentials give agent interactions a basis for trust and an audit trail of who authorized what.
Does this work with our existing AI stack?
Proven AI is built on open verifiable credential standards and is designed to add an identity and authorization layer to agents you are already building or deploying, through APIs and SDKs, rather than requiring you to adopt a particular agent framework.
Is this only for autonomous agents?
No. The same approach applies wherever identity and authority need to be established in AI interactions, including agents acting for a person, agents acting between organizations, and cases where a human needs to confirm they are dealing with a legitimate agent rather than an impersonator.
Next step
Give your AI agents an identity they can prove.
Customer chatbots, internal copilots, or autonomous agent networks: bring your architecture and we’ll show you how Proven AI integrates.
Book a session →
