We’ve just released a sneak peak of Ploy’s brand new agentic AI Ava. Think of Ava as having an extra employee on your team dedicated to ensuring every single identity is managed across your entire organisation, without having to lift a finger.

Ava will be able to:

• Automatically onboard and offboard users into the right tools and resources

• Simplify access reviews and provide more detailed context to reviewers to help them make more informed decisions

• Enforce company policies and remediate violations

• Proactively suggest policies and access conditions to ensure your security evolves as your business does

• Investigate and remediate misconfigurations across your integrations

• Execute ‘playbooks’ or common pre-defined workflows your team currently spend hours per weeks on

• ??? ← Your own ideas for Ava

This release is just a sneak peak that help you diagnose errors in your flows, helping you understand what’s gone wrong and more importantly how to fix it. Ava has been hooked up to your integrations so can also easily query against your live data to solve common questions like “why can’t I add xyz@acme.co to group ABC”?

While this is a beta release there a couple of limitations, most importantly requiring you to rate each conversation you have with Ava to ensure we have the feedback needed to make Ava the best identity AI assistant it can be. To get started with Ava, navigate to an error on flows and click the little ‘Investigate with Ava’ button:

A note on security:

We’ve spent a lot of time adding in guardrails and security features to ensure that Ava is protected against common attack vectors and ensuring it only has access to your data. Additionally nothing you send to or interact with Ava is ever used to train models or seen by model providers. All inference is provided by AWS Bedrock currently in the EU region.