How Stripe Built a new Internal AI Knowledge Platform that their PMs use “all day long”
Non-engineers are embracing internal AI tools. Examples from Stripe, Spotify, Uber and more.
🔒DoP Deep goes deeper into the concepts and ideas that are covered in the Weekly Briefing to help you learn lessons from the experiences of top tech companies. If you’d like to upgrade to receive these in-depth pieces of analysis you can upgrade below. New reports are added every month.
After Stripe unveiled its internal agentic coding suite called Minions earlier this year, it has since gone on to spur several product development teams to build agent coding fleets of their own.
Inspired by the results they saw with Minions, Stripe has now revealed a new follow up internal AI tool which has been designed from the ground up to explicitly target non-engineers. And according to a Senior PM at the company, this new tool is something he uses “all day long”:
And its not just Stripe who are refocusing their internal AI tooling efforts towards non-engineers. Uber’s CTO, for example, recently confirmed that they’re now deploying small groups of engineers (what he calls “Agentic Pods”) throughout their company to work with non-technical parts of the company to understand use cases and workflows - and then build custom-built tools around them.
In this Deep Dive, we’re going to unpack Stripe’s new Knowledge AI Platform (that it internally calls “Kai”) to understand what it does, the core technologies and guiding principles involved - and how you might be able to take some of these principles to build your own version of an internal AI knowledge tool that can be used for non-engineers.
As well as this, we’ll also cover other recent examples of similar AI internal tools and agents that were recently released from companies including Spotify, Uber, Intercom and more.
Coming up
What is Stripe’s Internal AI Knowledge Platform and how does it work? The core parts of the application unpacked
How you can build your own version of something similar - sample prompts, guidelines and workflows
Other examples of innovative new internal tools at world leading tech companies including Block, Asana, Uber, Rippling and more.
What is Stripe’s internal Knowledge AI Platform?
Here’s a snapshot of what Stripe’s new internal Knowledge AI platform is and how it works:
As you can see, the Knowledge AI platform is pretty expansive with its core capabilities surfaced across a number of different touch points including:
a web app
Slack integration
other embedded internal tools
and even a Chrome extension which is a super smart way to allow users to use it wherever they are.
This design means that one user might use the web interface to start an in-depth conversation (long context windows are baked in with one user reaching 932 turns) and another use might rely on the Chrome extension to plug into an existing internal tool without using the web UI very much at all.
Stripe says that this was a core architectural decision to make the tool as “surface-agnostic” as possible. The agent is built as a service rather than a single application and users can interact with it through any of its supported surfaces, though Stripe admits that the most common view is through the dedicated web app.
A closer look at what exactly it can do and the technologies powering it
Stripe says that the platform is used across many different departments within the company - and has shared some of the use cases and core achievements across them:
Sales Research - researching accounts before calls, generating 2x sales activity, creating 17% more opportunities
Data Analysis - 5,000+ daily sessions are centered on this alone - querying data warehouses, analyzing logs, surfacing metrics