Benchling launches Science Foundries: agent teams take over the handoffs from design to order to bench to analysis
Synopsis
Benchling released Science Foundries, a way for scientists and AI agents to divide work over the same scientific data and context and run processes such as protein production, lead optimization or process improvement end to end. A user picks a process or goal, then sets which steps agents run and where scientists must approve; agents execute and monitor work, coordinate handoffs and surface key decisions. In a high-throughput antibody example, design, cloning, lab and analysis agents connect to models, a CRO, lab automation and training data, so assay results finished late at night are analyzed before the scientist starts the day.
Interpretation
Scientific processes move from scientists manually carrying data between steps to agent teams that execute, monitor and coordinate handoffs on one shared body of scientific data and context. Previously scientists processed data by hand and fed it from one step to the next; a team might already have sequence-design models, a CRO synthesizing constructs, and lab automation for production and assays, yet the links between them stayed manual. In the high-throughput antibody design, production and testing example, a design agent generates and evaluates sequences, a cloning agent drafts the assembly strategy and places CRO orders, a lab agent dispatches workcell instructions, captures data and performs QC, and an analysis agent interprets results and updates training datasets.
A Foundry is built in two steps, choosing the goal and then assigning who does what, with agents executing and monitoring while scientists keep approval over key decisions. Agents previously handled discrete tasks such as defining candidate-target profiles, experimental design and SPR analysis; they are now organized across whole processes such as protein production, screening and lead optimization. Parabilis Medicines, a clinical-stage biopharma with a lean team running several programs in parallel, moved from chat to standalone agents to Science Foundries, as described first-hand by scientific co-founder and Head of Platform John McGee.
The whitespace between steps shrinks, so experiments can start sooner. When an assay finished at night, results waited for the scientist the next day; now agents pick the results up automatically and present an analysis by the time the scientist starts work. By the example's account, each experiment starting a day sooner can add up to an R&D program speeding up by months.
Perspective
The scope is Benchling's launch description of Science Foundries: the two-step build, the four agent roles in the high-throughput antibody process, and Parabilis Medicines' first-party account of moving from standalone agents to Science Foundries. It is aimed at researchers and engineers considering agents inside R&D workflows; teams already on Benchling can, as described, pick one process among protein production, screening, assay development, method development or in vivo work to build a first Science Foundry.
The claims that a three-person team can have the capacity of thirty and that an R&D program can speed up by months are the publisher's assertions without attached measurements; the Parabilis Medicines assessment comes from the company's scientific co-founder and is a first-party statement. The build lists for therapeutic area, preclinical and process development work are offered directions whose results are not reported.