Multi-Shot Conversational AI Interface for Identity Graph

Read full details here: https://www.linkedin.com/posts/vincenzoiozzo_back-from-the-vegas-madness-our-team-has-ugcPost-7360664976404979713-tMnD?source=nhimg

Fresh from Las Vegas, our team has launched a multi-shot conversational AI interface for the identity graph, enabling security and governance teams to ask complex identity and access questions in plain language, no query syntax required.

This release is designed for GRC, IAM, and Disaster Recovery (DR) teams who need deep visibility into identity relationships without spending hours crafting and testing queries. With multi-shot interactions, the system supports follow-up questions, context-aware filtering, and iterative refinement, allowing users to drill deeper without starting over.

Real-World Examples:

• Identify all users without MFA who can access a specific Snowflake table.

• List Okta accounts capable of federating into an AWS role with S3 access.

• Find Active Directory service accounts with credentials older than 90 days.

Why It Matters

By replacing traditional query-language barriers with a conversational approach, teams can save hundreds of hours and uncover identity risks more quickly. This is especially valuable in complex environments where identity sprawl, cross-platform access, and dormant accounts can create blind spots.

Key Takeaways from Development:

1. MCP Limitations - The MCP protocol alone isn’t enough for advanced, cross-system queries.

2. Runtime Awareness - Estimating query execution time improves both performance and user experience.

3. Multi-Shot Advantage - Iterative, context-aware conversations dramatically outperform single-query interactions for complex identity investigations.

Bottom Line

This update transforms the identity graph from a static data source into an interactive investigation partner, empowering security teams to quickly ask, refine, and act on identity intelligence without touching a line of code.