Savvy Savi?

Remember this day, folks. Saviynt fires the first shot in the AI Wars. Yesterday, Saviynt introduced its Intelligence platform to the world. There were some hits and misses ( in my opinion), and I’ll post a link to my deep dive on it, but suffice it to say, we got a chance to see their vision for AI in their platform. In a nutshell, they are leaning into the position of being the central place to not only administer identity but to secure it and analyze it. They unveiled an Identity Posture management module, certification recommendations, access request recommendations, and their AI named Savi.

What I liked:

• They leaned into the power of having an LLM and allowed natural language and conversational style interactions between the user and the platform. 👍🏿

• Looking to provide context into IAM administrative actions. How is automated provisioning affecting the business? How is the platform helping to reduce risk? 👍🏿

What I didn’t like:

• Certification Recommendations. More so, this feature is leading as a top-level feature. This capability has been around since 2018. Maybe not in their platform but this is using advanced technology to solve a really easy problem. It’s important because you want to reduce time here, but this is a super low-hanging fruit problem, given where we are today with technology. 👎🏿

• Don’t just recommend..automate. They pushed recommendation features on both access request and certifications. Recommending is cool, but you know what’s even better…Automating. They Missed an opportunity to create some distance here between themselves and the market. 👎🏿

With Oktane and Navigate right around the corner it will be interesting to see if we get any more news around “AI Enabled features”. I’m expecting Co-Pilots and AI jargon galore the next two quarters from all the major IAM platform players, and contenders.

Oh this will be so much fun!

Identity AI

The next decade will be one of the most transformational in human history.

We will completely rethink the way we interact with computers. I 1000% percent believe that, and you’re more than welcome to save this and send it to me 10 years from now if I’m wrong. But what I experienced this week, and the momentum I’m seeing with AI technology, I don’t think I’m wrong.

Let’s hop in our time machine real quick and go back to the year 1985…just kidding, let’s go to the year 2008. I’m in the movie theater geeking out to the first Iron Man and this scene happens.

I remember telling myself. That’s it! That’s the future. I want a real, live, interactive AI system that helps me build things faster. Granted, I realized I was looking at a movie, but the thought was captivating. So much it inspired my first foray into entrepreneurship, but that’s a story for a different day.

J.A.R.V.I.S ( Just A Rather Very Intelligent System) wasn’t designed to replace Tony, it was designed to help Tony. To speed up Tony’s ability to turn a thought into a reality. Mundane tasks like setting up project files, storage space, and connecting backed infrastructure let JARVIS handle that, while Tony focused on imagination. From that day forward, I’ve always dreamed of the day that JARVIS would be a reality and not a fantasy.

Back to the present…a couple of weeks ago, I started using OpenAI’s voice prompt ChatGPT. SPOILER ALERT: No, it’s not JARVIS, but the experience was close. I could just ask it questions, and it responded, even with some inflection in the vocal responses. I immediately thought, wow, I wonder how close we are to building JARVIS.

This got me to thinking about what it would mean to truly integrate AI into an IAM platform. Not just better mous trap features, but next level evolution features. Here’s what I came up with:

Table stakes Features:

Interactive Chat - This is baked in throughout the platform and understands the context of wherever it’s initiated. Backed by not just knowledge of the platform, but IAM knowledge.

Certification Recommendations - Recommendations for which entitlements to certify

Smart Access Request - Tailored access requests based on peer analysis, entitlement activity, and risk.

Automated Role Lifecycle Management - Use data analysis to create and manage roles based on usage.

Next Level Features:

Conversation-based configuration changes - A user can give commands at the next level of interactive chat instead of just querying for information. “Restrict all access to financial applications to only employees in the finance department.”

Automated Access Management - Access adjustment based on weighted risk score, feed by activity, and peer analysis. Enforces least privilege and continuously adjusts access.

Conversation-based reporting - Transforming IAM data into business insights. Type in a prompt and get reports. Bonus points for making it a voice-enabled.

Are these things possible? I believe so; I’ve spent the past week building prototype apps building UIs, all with prompt-based interfaces and AI-generated code. Took the development process from days to hours. More about that later, the point is AI is here to stay, and I challenge companies to lean into and not be afraid to innovate. This is the chance to take leaps when it comes to functionality. Be bold.