Overview
After deciding to pivot Bloom into AI, my co-founder came back with a specific idea he had been researching. AI agents were everywhere in 2025, basically software that can take actions and make decisions on its own, and the tooling to actually monitor what they were doing had not caught up yet.
The comparison he kept making was to Datadog, a tool companies use to watch over their servers and catch problems before they spiral. Nobody had really done that for AI agents yet. That felt like a real opportunity, and I was in.
Contribution
Product + Engineering
Team
Ruthvik Jonna
Aditya Naidu
Date
July 2025
Process
We moved fast, the way we always did inside Bloom. Cold outreach, customer interviews, and an MVP all running at the same time. The feedback we got was genuinely interesting. Most people agreed the problem was real. AI agents were multiplying inside companies and nobody had a clean way to see what they were all doing.
The pushback was consistent too. A Senior Vice President at SAP put it most directly. He said the space was getting crowded fast, and that we should think about what was adjacent rather than going straight at it. He thought we were onto something, just not quite the right angle yet. That kind of honest feedback is rare, and we took it seriously.
Outcome
That conversation stuck with me. We had already built a solid foundation and the technical work was not going to waste. But on the business side, we knew we needed to shift.
The pivot this time was softer than the ones before it inside Bloom. We were not throwing everything out. We were looking one step to the left. And what we found there ended up being the most interesting idea we had worked on yet.