TechCrunch Coverage: techcrunch.com/2024/10/03/kapa

At kapa.ai, we are on a mission to answer all technical product questions in the world. We turn our customers’ technical documentation into intelligent assistants which instantly answer their end users’ complex questions. We are proud to count more than 100 leading companies as customers incl. Docker, OpenAI, Monday.com, Grafana, CircleCI, Silicon Labs, TomTom, and Mapbox.

Today, we are thrilled to announce that we raised a $3.2M seed round led by Initialized Capital, with participation from Y Combinator and an impressive roster of angel investors including Douwe Kiela (Founder of ContextualAI, Stanford Professor, Author of original RAG paper), Amjad Masad (Founder of Replit), Solomon Hykes (Founder of Docker), and more.

Docker's AI assistant, powered by kapa.ai, provides instant, accurate answers to technical questions directly on their website, significantly reducing support ticket volume and improving user experience.

From Academic Research to Solving Real-World Problems

Back in 2019, during our Computer Science postgrad studies at Imperial College London, we were already exploring the potential of LLMs and early embeddings for information retrieval and question-answering systems. Little did we know how directly this research would apply to solving real-world problems just a few years later.

Fast forward to January 2023, and two of our friends running tech companies independently reached out with the same problem: their support teams were swamped with repetitive technical questions, despite having comprehensive documentation. This highlighted a need for faster, more efficient access to information.

Building Out A State-of-The-Art AI Product

Over the last two years, our team has built a comprehensive platform that solves a critical problem: transforming company knowledge bases into reliable, AI-powered assistants that instantly answer complex technical questions. kapa.ai has quickly gained the trust of over 100 leading companies across various industries, from cutting-edge startups to established enterprises.

The kapa.ai platform addresses every aspect of creating and maintaining a modern LLM-based AI assistant:

• Explainable answers: Minimal hallucinations with answers grounded in source materials, rigorously tested across 100+ deployments.

• Comprehensive knowledge integrations: The platform offers 30+ data source integrations (e.g., website crawlers, Slack channel readers, wiki/forum parsers) that automatically refresh, ensuring up-to-date information.

• Off-the-shelf deployments: 1-click deployments including website chatbot widgets, Slack bots, Discord bots, Zendesk assistants, and APIs for easy integration.

• Continuous improvement: Built-in features and analytics help find gaps in your documentation, adjust AI behavior, and enhance answer quality over time.

• Enterprise-grade security: SOC II Type II certified with advanced PII detection/masking technology for data security and privacy.

Building customer trust in AI requires a robust and collaborative platform that companies can effectively work with to continuously improve their AI assistant's performance and knowledge base.

Becoming World-Class at LLM Evaluation

Running a retrieval-augmented-generation (RAG) application in production we quickly realized that our success hinges on our ability to measure its performance. It is impossible to optimize something you cannot measure. As a team we have spent the last year becoming experts in this field. Academic benchmarks and leaderboards only take you so far. Our success is based on our ability to quantify how good a given version of kapa is at answering technical questions. A large part of our engineering efforts are spent incorporating customer feedback and usage patterns into our evaluations. This foundational work allows us to reliably test and incorporate new research into our production systems.

Douwe Kiela, Stanford Professor (author of the original RAG paper), Founder of ContextualAI, and an investor in kapa.ai said this:

“The kapa.ai team stands out for their thorough and novel evaluations of LLM-based question-answering systems. By constantly testing all permutations of the latest models and academic techniques, they stay ahead in providing trusted solutions to their customers.”

Real-World Impact Across 100+ Customers

We have been amazed by how quickly we have grown. We are now working with over 100 leading companies, from startups to large enterprises, and it is incredible to see kapa.ai becoming a trusted part of their product experience. What excites us most is the real impact we are having. For example, Mapbox reported a 20% monthly reduction in support tickets thanks to improved self-service. Docker told us they are now helping over 20,000 Docker users find instant answers through their AI-powered Docs Assistant. And CircleCI shared that they have cut their response times by 28% by empowering their technical support agents with instant AI-drafted responses. These are not just numbers to us - they represent real people spending less time stuck on technical issues and more time being productive.

“As a developer, I’ve spent countless hours picking through hard to navigate documentation and find myself not really knowing where to start looking for what I need on large knowledge bases,” says Brett Gibson, managing partner at Initialized. “The utility of Kapa was immediately apparent, and it’s no wonder they’ve quickly become the go-to tool for some of the biggest logos in tech. Emil and Finn not only understand how to build a product that gets developers the answers they need, they also deeply get developer relations and community engagement.”

The Road Ahead

With this funding, we continue to be well positioned to further enhance our AI platform, expand our integrations, and continue to improve the product experience for our customers across various industries. We are actively recruiting talented individuals who are passionate about building tools to help companies provide exceptional technical product experiences. If you are interested in joining our team, check out our open positions.

If you are interested in seeing how kapa.ai can transform your knowledge base into an intelligent assistant, you can test it out by requesting a demo here.

Our goal is simple: enable your users and employees to find quick, accurate answers to complex technical questions. This means less time troubleshooting and more time using your products effectively. We're focused on building AI-powered solutions that make this a reality.