Company
Enjamb is backed by Y Combinator
Y Combinator is backing Enjamb as we build domain-specific AI for the teams running drug programs.
by Enjamb Team
Today we are announcing that Enjamb is backed by Y Combinator.
Enjamb is domain-specific AI for biopharma. Our agents work inside the systems a drug program already runs on, across research, preclinical, clinical, regulatory, and quality, and every result they return stays attached to what produced it.
We built it this way because the bottleneck in drug development is not a shortage of good questions. It is the distance between a question and an answer somebody is willing to sign their name to.
The work is spread across a stack nobody chose
One program touches an electronic lab notebook, a LIMS, a clinical data platform, a regulatory document system, an object store, a statistical environment, and a decade of internal tools that exist nowhere else in the world. None of them were built to talk to each other. They were bought at different times, by different functions, under different constraints, and every one of them is load-bearing.
So the work between them gets done by people. A scientist exports a dataset, reshapes it, runs the analysis somewhere else, pastes the result into a document, and hands it to a reviewer who has to reconstruct the whole path before approving anything. The expertise in that chain is real. The hours spent carrying material between systems are not where it should go.
General-purpose AI does not close that gap, because the hard part was never the reasoning. The hard part is reaching the systems, holding the context across a task that runs for days, and producing something an accountable person can actually check.
Agents that work inside the stack you already run
Enjamb agents do not ask a team to move its data somewhere new. They connect to the systems the work already lives in and operate there, under the identity of the person who asked, with that person's existing permissions.
That is the difference between an assistant and a colleague. An assistant needs everything handed to it. A colleague goes and gets it.
- Research systems, including electronic lab notebooks, registries, and screening platforms
- Clinical and regulatory systems, including clinical data platforms and controlled document vaults
- Storage and compute, including object stores, notebooks, and the scripts that produce the deliverable
- The everyday layer, including document and spreadsheet editing, email, and internal wikis
- Your own internal tools, through the same connector model rather than a bespoke integration project
Context that survives the session
Most AI products forget everything the moment a conversation ends. That is survivable for a one-off question and useless for a program that runs for years.
Enjamb keeps a shared working context across a company: the decisions, the analyses, the tradeoffs that were considered and rejected, the internal conventions that never made it into a document. Agents read from it and write back to it, so the second person to ask a question starts where the first one finished rather than at zero.
It is scoped the way a company is scoped. Something written for one program stays with that program. Something that belongs to the organization is available to the organization. Nobody sees anything their permissions did not already give them.
Every result keeps its sources
In biopharma an answer without provenance is not an answer. It is a claim somebody now has to go and verify, which is most of the work you were trying to save.
Every Enjamb result stays connected to what produced it: the documents it read, the systems it queried, the code it ran, the assumptions it made, and the person whose request set it going. Consequential actions stop and wait for a human. The reviewer sees the path, not just the conclusion.
Teams should be able to delegate more of the work without delegating away the judgment. That constraint shapes every part of the product.
Where the work already happens
The other thing we believe is that good software meets people where they are. A scientist should not have to open a new tab to hand off a task.
Enjamb runs where the conversation already runs, in Slack, Microsoft Teams, and Google Chat, as well as in the workspace itself. Ask in the channel, and the run comes back with its sources attached.
What comes next
Y Combinator's backing gives us more room to build alongside the research, clinical, regulatory, and quality teams doing this work every day. We are focused on the workflows where extra capacity compounds and where traceability cannot be bolted on afterwards.
If you run one of those teams, or you want to build this with us, we would like to hear from you.
