# Leaders at Glean must build their own priority agents

Markdown copy of https://realalpha.app/s/leaders-at-glean-must-build-their-own-priority-agents/ on Real Alpha, the best minutes from long interviews with the people building AI. Cite and link the page, not this copy.

- Speaker: [Arvind Jain](https://realalpha.app/builders/arvind-jain/) · Arvind Jain is founder and CEO of Glean. · Founded Glean, the AI work assistant, after co-founding and leading R&D at Rubrik and a decade at Google.
- Host: [Nakul Mandan](https://realalpha.app/people/nakul-mandan/) · Host, Knuckle Up with Nakul · General partner at Audacious Ventures; his show goes deep with founder CEOs on how they build.
- Show: Knuckle Up with Nakul
- Episode: [$7B AI founder: "AI is not capable of replacing you right now" | Arvind Jain (CEO, Glean)](https://realalpha.app/v/zPtwI6JaRy0/), Sep 9, 2026
- Moment: 33:11 to 36:49 of the episode (4 min), playing on https://realalpha.app/s/leaders-at-glean-must-build-their-own-priority-agents/; the episode from there: https://realalpha.app/v/zPtwI6JaRy0/?t=1991

## Quote

> “we're going to actually **ask our leaders to actually go and build** all of those agents”
> — [Arvind Jain](https://realalpha.app/builders/arvind-jain/) at [34:30](https://realalpha.app/v/zPtwI6JaRy0/?t=2070) · Knuckle Up with Nakul

## Takeaways

1. Agent Finder ranks work by time spent, not complexity.
2. Leaders must build priority agents, not delegate them.
3. They can use Glean or another agent platform.

## About this segment

Glean ranks automation candidates by where employees spend time, then asks leaders to build the agents themselves. Arvind Jain says they cannot delegate the work.

From “$7B AI founder: "AI is not capable of replacing you right now" | Arvind Jain (CEO, Glean)”, Sep 9, 2026. [Watch the full episode](https://realalpha.app/v/zPtwI6JaRy0/)

## Transcript

Transcribed by machine (Whisper) from the episode’s audio, so a word may be off here and there.

[33:16](https://realalpha.app/v/zPtwI6JaRy0/?t=1996) you know, which we just could not, you know, get done. At this point, is every department at Glean pretty AI natives? Like engineering is obvious, but like finance and back office, HR. So we are like, since we are one of the first enterprise AI products, so we've been using it heavily inside. And so finance team has, you know, tons and tons of agents. Like they've actually one of the most AI-fied department for us.

[33:41](https://realalpha.app/v/zPtwI6JaRy0/?t=2021) To be honest, you know, we are, I'm not satisfied with how much AI-ification we have done. you know we could we could do better last quarter i started a program so we have this this product inside lean called agent finder and agent finder is it uses that same sort of knowledge graph and understanding of like you know what work is getting done and who's doing it and actually it actually prioritizes and figures out that these are the the top 20 or 30 agents that you should actually build because this is where most of your human time is actually getting consumed today and it's rudimentary in the sense that it is basically basing it on where time is being spent we're not actually saying that hey like but this is too hard and therefore this should not be like it's just giving us where the time is being spent right

[34:25](https://realalpha.app/v/zPtwI6JaRy0/?t=2065) and so based on that we have a list of things that are not yet agentified what we're going to do is we're going to ask like not engineers and not individual contributors but we're going to actually ask our leaders to actually go and build all of those agents and they can build it in Glean or they can build it in any other agent platform it doesn't matter And so that's the program that we're running right now.

[34:46](https://realalpha.app/v/zPtwI6JaRy0/?t=2086) So I was trying to achieve two things with that. One is, of course, building more automation. But also a big part of how that's going to be achieved is when you truly make a workforce AI native, right? And so by forcing our managers to do it, and they cannot delegate it, that gets to a place where our leaders are fully AI-fied. Do you feel your non-engineering leaders are at this point AI-fied enough where they are getting into cloud code.

[35:14](https://realalpha.app/v/zPtwI6JaRy0/?t=2114) Because cloud code is literally writing the code itself. So you don't even need, as long as you can be a product manager, you could probably build agents yourself. I know that's outside of Glean, but do you feel you've kind of, through the brute force of last couple of years, made that a mandate? If you're not AI native as a manager, can you survive at Glean? Let me ask you this way. We have a process, actually.

[35:36](https://realalpha.app/v/zPtwI6JaRy0/?t=2136) Even when we recruit people, there is the AI native check that happens. if today like you know after all like ai is already very you know uh quite spread you know well spread right so in our interview process we we actually test for how much you've been using cloud and chat gpt and gemini and what work have you done with it in the past most of our managers You don't build agents in Cloud Code. Our engineers, of course, use Cloud Code all the time to write code because, you know, like Glean is simpler and it's much easier to actually go and build these agents in Glean.

[36:13](https://realalpha.app/v/zPtwI6JaRy0/?t=2173) And so our, and that's the interesting thing, like, you know, if you think about it, while, you know, AI is a very powerful technology and everybody should know how to use it, but it should not feel like a complex technology that people have to learn. So I would still want our sales leader to be primarily very, very good at sales leadership. And yes, they should be aware of the power of AI, but we do believe that the tools will be quite easy for them to just use.

[36:40](https://realalpha.app/v/zPtwI6JaRy0/?t=2200) Like they don't have to actually become hackers or builders. There's been this dream of an AISDR.