This is part one of a three-part series on how HubSpot has transformed itself with AI. Part two illustrates how we grow with Agent-first GTM. The third part explains how we operate as an AI-first company.
Everything we build at HubSpot exists to help our customers grow. So when generative AI emerged, our engineering team didn’t just see a productivity tool; we saw the opportunity to create better products and put more value in the hands of customers faster.
And when standard AI tools reached their limits, we didn’t just look for better ones. We built the platform underneath them. That decision compounded faster than we expected. Because all of our AI is built on shared foundations, every new feature we deliver makes the entire system more powerful, and customers get a more consistent experience across everything they use.
Today we are able to innovate at a pace that was simply not possible before. 100% of our engineers use AI, and we’ve seen a 73% increase in lines of code written by our engineers.
We didn’t get here overnight. It took three phases, real infrastructure investments and the will to build what didn’t yet exist. Here’s how we did it.
Phase 1: Productivity with co-pilots (2023-2024)
In 2023, large language models had just crossed the threshold of being truly useful in a coding context. The best solution to using AI in engineering was to start with what was proven. At the time, it was about code completion: a human writes the code, and the AI co-pilots suggest what comes next.
We implemented a coding co-pilot and quickly achieved 30% adoption. Then we mined crash data, compared teams that used co-piloting to those that didn’t, and showed that adopting AI had no negative impact on product reliability.
With this data in hand, we removed the guardrails and gave everyone co-pilot access. Adoption surpassed 50% almost overnight. This taught us a lesson about how we make decisions. Measure, demonstrate and then scale.
By the end of Phase 1, 80% of engineers were using AI tools. We saw a 51% improvement in design velocity, meaning engineers shipped working code to production much faster, and a 7% increase in lines of code updated per engineer. We have proven that AI can make every engineer faster without compromising product reliability.
Phase 2: Scaling with Encoding Agents (2024-mid 2025)
The next step was autonomous coding with agents. Our teams could provide the tools to complete tasks end-to-end. Agents could read context, write code, run tests, and fix errors, all while the engineer reviewed and guided. We firmly believed that this was the future of engineering and were fully committed to it.
The real constraint came early. Standard coding agents couldn’t access internal build systems, our libraries, or verify that the code actually worked in our environment. Therefore, we built the agent integrations ourselves using MCP, a standard that allows AI agents to connect to external tools and systems, and deployed them to every engineer. To drive adoption, we hosted events to give engineers a dedicated space to learn, experiment, and become familiar with the new tools. Agent usage went from zero to 80% in one month.
The next challenge was the ladder. The engineers wanted multiple agents to run in parallel, overnight, without supervision. So we built an agent execution platform on top of our Kubernetes infrastructure. Each agent runs inside an isolated container that replicates a real HubSpot development environment. Agents compile code, run automated tests, read error outputs, and iterate on their own until everything works. No human intervention required.
By the end of Phase 2, 96% of engineers were using AI tools, design speed had increased by 60%, and lines of code updated per engineer had increased by 48%. We were starting to ship better products faster with agents. But that was only the beginning.
Phase 3: Scale with our AI platform (mid 2025-present)
HubSpot’s platform approach to product development has always been how we have created the most value for the customer. When we built reporting and automation at the platform level, we didn’t just provide a feature; we have deployed this feature to each hub at the same time. This is how innovation comes together.
We applied the same logic to our AI infrastructure in Phase 3. Instead of building each agent from scratch, we created the shared foundation once: how agents access data, what actions they can take, how they connect to the rest of HubSpot. Everything revolves on top of it.
The result is that all our agents are interoperable. They speak the same language, share the same tools and draw from the same context. A customer gets a consistent experience regardless of which agent they use because, underneath it all, they are all built on the same infrastructure. And because they’re all connected, every new feature we add makes the entire system more valuable. This is something that a collection of point solutions cannot replicate.
And it’s been made possible by the way we’ve scaled engineering with AI. Today, 100% of our engineers use AI, updated lines of code per engineer have increased by 73%, and the time to first feedback on pull requests has decreased by 90%. This means less time waiting and more time shipping what customers actually use.
Why it matters: Increase customer value
Having the right infrastructure accelerates the pace of innovation. For HubSpot, every agent we create makes the platform more powerful. Every piece of context we add to the platform makes every agent more effective. For customers, this means the product continues to get better, faster and more connected.
What used to take months now takes weeks, and those weeks translate directly into new capabilities in the hands of marketers trying to reach the right audiences, reps trying to close deals, and Customer Success Managers trying to build customer loyalty. They don’t need to think about the underlying platform. They can simply experience the result.
