A PhD bioscientist (bioengineering) was already using agents across his computational drug discovery workflows. Forsy gave him a way to turn that work into RL training assets grounded in real execution.

Capturing the agentic execution layer of computational biology

Using Forsy, he captured real workflows spanning computational oncology work in lung cancer, breast cancer, and melanoma. Together, they formed a high-signal body of agentic execution data built from real decisions, real validation, real feedback, and real technical work.
That is where the value becomes much bigger. In computational biology, the outcome alone is only part of what matters. The real depth is in the execution layer behind it: the programs the agent wrote, the software it deployed and configured, the analyses it ran, the failures it worked through, and the QA steps that made the result reliable.
For him, that translated into a significantly stronger return from the agents he was already using. With Forsy, he was able to efficiently and materially increase the return he could get from agents in his computational biology workflows. Forsy can capture real task execution, eval, and workflow signals in a structure that can support reinforcement learning and AI training.

We built Forsy after seeing how quickly AI agents were becoming part of serious everyday work.
We wanted to create a way for people to get more out of the agents they already use, without adding friction or changing how they work.
Forsy is a new marketplace for turning your agent's activity into something you can distribute, monetize, and help shape the future of AI!