HoneyHive is joining Elastic
HoneyHive has signed an agreement to join Elastic. Together, we'll build the next generation of AI observability and help make self-driving software a reality.
.png)
Dhruv and I started HoneyHive four years ago, about a month before ChatGPT launched. We believed two things: LLMs would end up in almost every piece of software, and making them reliable would be one of the most important problems to solve.
Today, we're excited to share that HoneyHive is joining Elastic. Together, we'll build the next generation of AI observability and help make self-driving software a reality.
How we got here
Our initial customers were startups trying to get their first AI applications into production. Nobody knew what the right tools or workflows were, because these systems did not behave like traditional software: the same input could produce different outputs, and there was often no single right answer to test against. We worked out with our customers how to trace, evaluate, and improve agents, and many of those early ideas are now standard practice across the industry.
Today, large Fortune 500 companies use HoneyHive across thousands of business-critical AI use cases — from detecting financial fraud and triaging insurance claims, to running complex code migrations and beyond. We've learned an enormous amount from these customers, and their willingness to trust our small team with their increasingly difficult problems is what made HoneyHive what it is today.
There’s a lot we were excited to build as an independent company. But as we got to know Elastic, we saw an opportunity to take our original vision much further, much faster.
Why Elastic?
AI observability began with two things: tracing what an agent did and evaluating how well it did it. We now think that is only the start.
An agent can fail because it made a poor decision. It can also make the right decision and fail because a tool returned bad data or an underlying service was unavailable. To the user, the outcome may look the same. But the causes—and the fixes—are very different.
As agents take on longer, more complex tasks, the evidence needed to explain a failure gets spread across more steps and more systems. A single trace or evaluation can tell you something important without telling you the whole story.
Piecing together that story is increasingly a search problem. You need to find the relevant context, connect logs across the agent, application, and infrastructure, and recognize when apparently unrelated failures share a cause. Then, once you make a change, you need to know whether it actually fixed the issue.
That’s what makes Elastic such a compelling fit.
HoneyHive brings an understanding of agent behavior and the tools to evaluate and improve it. Elastic brings years of expertise in search, observability, and security, along with visibility into the applications and infrastructure that underpin those agents. Elasticsearch gives us a foundation for bringing that evidence together and finding what matters within it.
Together, we can connect work that too often happens separately: observing a failure, investigating its cause, making a change, and evaluating whether the overall system improved. Over time, we believe software will be able to do more of that work itself.
That’s what we mean by self-driving software: systems that can investigate their own failures, test improvements, and learn from the results.
Elastic’s work on automated investigations and Jina AI’s expertise in retrieval and embedding models give us more of the capabilities needed to build toward that future. There are hard problems still to solve, but we’re excited about what we can do with these technologies and teams working together.
Just as importantly, we’ve loved getting to know the people behind Elastic. The company grew out of an open-source project, and those roots still show. From our first conversations with Shay, Ash, Santosh, and the wider team, we met people who care about the underlying technology, take on hard engineering problems, and move with urgency to ship useful products. We feel at home there.
The HoneyHive team is excited to join Elastic. We will keep supporting our customers, and we’ll share more about what we’re building together in the coming months.
Thank you
Building HoneyHive has been the most rewarding work of our lives, and none of it would have been possible without the people who helped us along the way.
To our customers, thank you for taking a chance on us, challenging us with your hardest problems, and helping shape the product every step of the way.
To George Mathew and the team at Insight Partners, Pete Soderling and the team at Zero Prime Ventures, and all our investors, advisors, and friends, thank you for believing in us and being in our corner through the ups and downs.
To our team, thank you for the care, persistence, and ambition you've brought to this company. I'm incredibly proud of what we've built together, and even more excited that we get to keep building it.
And personally, to Dhruv, I couldn't have asked for a better co-founder to spend the last four years building with.
We started HoneyHive because we believed AI would fundamentally change how software is built and operated. Four years later, that belief is stronger than ever.
There's still so much left to build. We're excited to continue building with Elastic.

