Episode 142 - GluuFederation/identerati-office-hours GitHub Wiki

Title: IGA: From automation to intelligence

Channels

Description

How will Identity Governance & Access (IGA) programs evolve beyond rote automation into adaptive, machine-learning-driven intelligence. Manoj will share insights from his recent articles on applying ML to access governance and on avoiding the “doomscrolling” traps that derail IGA initiatives. We'll discuss real-world patterns and anti-patterns, including how enterprise data models, contextual signals, and policy-driven decisioning can create smarter, more responsive governance. This session will offer actionable ideas to transcend compliance and deliver true intelligence in access control.

Homework

Event in London (Nov 25)

Takeaways

  • ⚡ The catalyst for IGA tools was regulatory compliance, not security. And perhaps, not-surprisingly, not much security benefit has been derived in the process--which is arduous and expensive.

  • ⚡ Ideally, governance should be risk-oriented. You need audit--i.e. to hold people accountable, and to detect threats. But preventing people from bad actions isn't the most effective way to ensure cybersecurity.

  • ⚡ Millions of permissions and thousands of applications--from a data science perspective, it's not a huge data set. For this size of data, graphs are a good fit.

  • ⚡ New tools for goverance are needed, and they won't look the same. But if they improve UX and security, it's bad news for the Sailpoints and Savyints of the world. Can the old dogs learn new tricks? Given the total lack of innovation in the last 17 years, don't hold your breath.

Cisco Article on "TBAC"

This article is very aligned with Manoj's vision:

The key is that the end user experience is good so the team feels good about the system!

Livestream Audio Archive

here