Emil

And we're live! Check out the @TechCrunch article here: techcrunch.com/2024/10/03/how… :)

user of @harnessio just spotted new kapa feature in the wild 👀

The docs: #scim-provisioning" target="_blank" rel="nofollow noopener">docs.kapa.ai/account/authen…

Yesterday was the most stressful day of my life. I ran straight to the nearest restaurant, jumped on a table, and shouted the most important words I'd ever shouted: "kapa NOW SUPPORTS SCIM PROVISIONING!" The closest waiter frantically responded "But Emil, what does that mean?" I then calmly explained that it means every Kapa customer can automatically provision and deprovision users through their identity provider. "Oh, so like if I use Okta, and I hire an employee, they automatically have a Kapa account?" "Yes!" I said to the surprisingly technical waiter. Then, I stepped down, paid for everyone's meal, and set up a @kapa_ai assistant for the restaurant, just in case they ever launched an API portal and technical docs. Did this happen? It could have. But we DID actually launch SCIM provisioning, and you should really check it out. Link is below. Now, back to my stress-free week.

Startup goal: design a product so good, it's just as good at answering questions as our customer success leader, and faster too 😂 (shoutout Simon, who is always around if you want to talk to a human about your @kapa_ai assistant)

The docs: #scim-provisioning" target="_blank" rel="nofollow noopener">docs.kapa.ai/account/authen…

Yesterday was the most stressful day of my life. I ran straight to the nearest restaurant, jumped on a table, and shouted the most important words I'd ever shouted: "kapa NOW SUPPORTS SCIM PROVISIONING!" The closest waiter frantically responded "But Emil, what does that mean?" I then calmly explained that it means every Kapa customer can automatically provision and deprovision users through their identity provider. "Oh, so like if I use Okta, and I hire an employee, they automatically have a Kapa account?" "Yes!" I said to the surprisingly technical waiter. Then, I stepped down, paid for everyone's meal, and set up a @kapa_ai assistant for the restaurant, just in case they ever launched an API portal and technical docs. Did this happen? It could have. But we DID actually launch SCIM provisioning, and you should really check it out. Link is below. Now, back to my stress-free week.

@bukovinski If only we had a shared Slack channel 👀 - see you in there!

@emilsnotes We should talk, would love your feedback. And we have a sneak peek of something a bit more powerful available for you. github.com/PSPDFKit/pdf-t…

Get started here: docs.kapa.ai/dev/agent/

You've probably noticed every app looks like this now. Let me explain why, and why chat isn't the real breakthrough. Chat was the natural first interface for LLMs. Easy for AI to understand, easy for users to interact with. So, naturally, as ChatGPT / Anthropic / etc crossed billions in revenue and hundreds of millions of users, most other apps followed suit. Even our experience is chat-first. 200+ customers with hundreds of millions of users, all asking techical product questions through AI chat. But what they really wanted was the outcome! Think about it. No one logs into Salesforce to use Salesforce...they log in to grow their business. We saw the same usage pattern across Kapa. Chats like "How do I integrate Google Drive?" are actually asking "Can you integrate my Google Drive instance for me?" This is where the puck is heading, and why companies like PostHog, Linear, and Notion are following in Claude / ChatGPT's footsteps. Believe it or not, building this infrastructure is brutal. To achieve this, you must focus on what happens after the chat request. Harness engineering, real-time RAG, orchestration, user-scoped auth, PII handling... But that's the user unlock. An in-product assistant. They don't have to leave, and the assistant has access to your docs, their tools, and knows how to navigate both. That's what the @kapa_ai Agent Framework finally does. The Kapa Agent Framework. Give it your tools and actions, connect one auth endpoint, and your users get a full in-product agent. Port, Matillion, and Medusa already use kapa as part of their product agents, and we're onboarding more Kapa customers every day.

@cncf @kapa_ai Read more here: contribute.cncf.io/blog/2026/04/0…

I'm extremely proud to announce our partnership with the Cloud Native Computing Foundation ( @CNCF ). Just as proud to announce our donation of @kapa_ai to any CNCF project. If you're not aware of the CNCF, you're surely aware of the projects it supports. Kubernetes, Prometheus, Envoy...all open-source, all maintained by the CNCF. This foundation spun out of the Linux foundation in 2015, and It is arguably the most influential open-source foundation in modern infrastructure. We ❤️ open-source, too. Kapa already powers the AI docs experience for huge projects like Opentelemetry, OpenFGA, and Envoy Gateway. That's why this partnership is so huge. It expands access for any non-commercial CNCF project. If your project is part of the CNCF, you can apply for free access directly through the CNCF Service Desk. There's a guide below. You'll get free access to the only CNCF-approved AI docs assistant. A lot of smart people agree it's a helpful AI assistant, and our customers like OpenAI, Silicon Labs, and Nokia would say the same. We're proud to be the only docs AI assistant approved by the CNCF. Hit the link below to learn more, and long open-source!

we've spent 3 years making sure AI can answer hard technical questions without hallucinating. 200+ companies trust us for this - Grafana, OpenAI, Nokia, Monday. last november, tool calling took a leap. we saw the opening: go beyond chat. we spent months with 40+ customers watching them try to build in-product agents from scratch. same story every time: months of infra, ship a v1 without evals, nobody maintains it, new harness drops. they all fall behind. so we solved it. you can now give @kapa_ai your tools and actions, connect one auth endpoint, and ship a real in-product agent.

Lots of companies deploying really, really bad and unhelpful AI agents. Meanwhile, Logitech quietly launches an AI assistant that's 99% accurate. A year later, tickets per unit sold are down 10%, help center content grew 30%, and sales starts using it to answer technical questions on calls. When AI works, people actually use it. When it doesn't, trust evaporates.

"We need to clean up our docs before we launch the AI assistant." This is the easiest way to waste 6-12 months. And I say that as a huge proponent of clean docs. AI has gotten so good, if you point it at your docs, it tells you exactly what to fix. That's because users will tell you. They'll talk to the AI, and the AI will surface questions it can't answer. You'll learn more from a day of user questions than a year of "improving the docs." Trust me. Companies don't improve docs and add AI. They add AI, and that improves their docs.

Talked to a Fortune 500 tech company last week that's desperately trying to pull their users AWAY from ChatGPT and towards their official docs. It's an uphill battle. The harsh truth is most engineers ask an LLM before visiting your docs page. But that creates a vicious cycle. 1. Customers ask ChatGPT instead of the docs 2. ChatGPT has outdated, unreliable, generalized information 3. Customers lose trust in the company because they never get the right answers I told the F500 company the only way to break out is to create a better AI experience. An AI trained on your product knowledge is a week or two away, maximum (if you partner with @kapa_ai) and it builds actual trust. Your users are already using AI to learn your product. The question is whether they're using one that actually knows your product, or one that's guessing based on two-year-old training data.

Last week was special for @kapa_ai. GSMA Insights demoed our custom AI assistant, live on the demo floor, to industry-leading companies like Google, Nokia, Ericsson, Qualcomm, and Samsung. You may have heard of MWC, and GSMA is the reason you see incredible new launches out of Barcelona every year. Every attendee is grateful for their hard work, ourselves included. They do more than put on the best conference on the planet, though. They make industry-leading reports on the global mobile ecosystem. Those reports hold insanely valuable (and technical) insights. We created a custom Kapa AI assistant on top of those reports, so anyone can ask questions and get the insights. It was a huge honor to see GSMA demo this AI assistant to their customers, and see the reaction from the conference. Thanks to Matthew Iji and Emanuel Kolta for letting us show you how great an AI assistant trained on GSMA reports could be. Can't wait for next year!

Have you ever noticed out-of-the-box LLMs almost never say "I don't know"? To me, that's a huge red flag if you (or your customers) are using it for production code. It confidently gives you something that sounds right but will break your production environment. A networking hardware company told me this when they were evaluating AI assistants: "If the answer is not in the documentation, the answer is made up and is plausible. Even if not necessarily correct." Sums it up perfectly. "Plausible" doesn't cut it when customers are associating an LLM's answer with your brand.

"AGI is right around the corner" I asked ChatGPT to transcribe a voice message, it said it couldn't do it. I asked Claude to do it, it tried to install OpenAI. We're still early.

A fintech company handling 32,000+ documentation questions per week described their current AI search solution as "fairly amateur." That's why it fell flat. "Technically works" isn't going to cut it for banking operations, when a hallucination means a compliance issue. Generic LLM pointed at complex product docs will never work for this reason. The LLM needs to actually understand the semantic structure and content of the docs!