Each Workflow Agent is purpose-built for a recurring deliverable that pharma and biotech teams produce week after week. Cheiron comes with 50+ agents organized by function:
RA: regulatory pathway comparisons, precedent searches, guidance summaries
Clinical: trial precedent analyses, endpoint benchmarking, study design comparisons
Medical Affairs: TPP drafts, response document drafts, advisory board briefings
CMC: manufacturing precedent searches, quality guideline summaries
QA/QC: inspection precedent searches, SOP gap analyses
R&D: target validation dossiers, mechanism reviews
BD/IP: competitive intelligence briefs, pipeline scans, patent landscape summaries
Commercial: launch performance comparisons, market landscape briefs
PV: safety signal investigations, post-market adverse event summaries
Each agent runs a specialized literature scan, applies the analysis pattern that deliverable typically requires, and returns a structured draft you can edit and ship. Like everything else in Cheiron, every claim in the output is cited and verifiable through Deep Citations.
Three ways to use an agent
From the home page
When you already know the deliverable you want to produce, go straight to the Workflow Agent section below the input box on the home page.
Filter by function: All, RA, Clinical, MA, CMC, QA/QC, R&D, BD/IP, Commercial, PV
Pick an agent: each card describes what the agent produces
Choose how to start:
Use Agent: enter your own inputs (target, indication, comparator, year range) and run the agent against your specific question
Try Example: generate a draft from a pre-filled example. Useful when you want to see what an agent produces before committing to your own inputs.
Run: Cheiron does the literature scan and returns a structured draft
From the query box
When you're typing a question and know an agent would handle it better, click Agents inside the query box to attach one inline.
From a conversation (follow-up suggestions)
When you're already in a conversation with Cheiron and an agent would help carry the work forward, Cheiron surfaces one inline as a Follow-up suggestion.
Click the suggestion and the agent launches with your conversation context pre-filled — the question, topic, and relevant entities are already populated. Edit the inputs if you need to, then run.
This means you can move from "I have a question" to "I have a structured deliverable" without bouncing back to the home page or rebuilding your inputs.
Note: Not every conversation gets an agent suggestion. Cheiron only proposes one when an agent would meaningfully extend what you've been working on.
When to use a Workflow Agent
Use a Workflow Agent when:
You need a structured deliverable (a TPP section, a competitive brief, a regulatory comparison) rather than an open-ended answer
You're producing the same kind of output regularly — agents save the most time on recurring work
You have specific inputs (a target, an indication, a comparator) and want consistent structure across drafts
You want a draft to edit rather than raw research
Note: Use Ask Cheiron when you have an open-ended question. Use Document Chat when your question is about a specific source. Use a Workflow Agent when you have a deliverable to produce.
Future Enhancements
Cheiron continues to improve Workflow Agents by expanding long-context processing and multi-generation capabilities, enabling agents to handle more complex workflows and longer inputs.
Future updates will support generating structured drafts for outputs such as CSR documents, IND sections, regulatory documents, and strategic reports.