FALCON Verify is launched as a new AI verification capability that analyzes AI-generated responses inside ChatGPT and Claude to determine what is supported, unsupported, contradicted, or insufficient based on available evidence.

The verification process preserves the existing user workflow by not requiring prompt rebuilding or content transfer to external tools, ensuring a seamless addition of a verification step.

To tackle AI trust issues, FALCON Verify shows the evidence behind each claim instead of relying on a single confidence score and creates a shareable verification record with a unique ID for traceability.

Usage workflow involves asking the AI, then saying “FALCON, verify that,” after which FALCON analyzes the response statement by statement against evidence and flags each claim.

FALCON Verify delivers quick results (under a minute), offers a free plan with 25 checks per month, and can be added as a connector to ChatGPT, Claude, or integrated into software and AI agents.

FALCON Verify is part of 1-i.ai’s broader AI Control Plane, which aims to help organizations govern and operate AI across models and providers, advancing robust, reliable, and responsible AI infrastructure.

The product provides examples showing how it identifies supported, unsupported, contradicted, and insufficient statements, helping users focus on attention-worthy claims.

The announcement frames FALCON Verify as a solution to the AI’s tendency to be confidently wrong, citing studies and internal tests about the prevalence of incorrect or unsupported statements.