Answers “how is the work done?” One canvas: humans and agents work the same document as co-contributors — drafting, challenging, reviewing in one flow.
It behaves like a modern editor, with one difference: the Chief of Staff and the expert services work right where you write. Ask, Consult, Job and Deliverable results live on the page as embeds that show live status and trace to their sources.
Headings, callouts with note / warning / limited-evidence variants, native tables with score columns, references and a provenance footer.
Summarize, Takeaways, Fact check — plus pharma actions: tighten regulatory phrasing, add HEOR framing, add a graded citation, add a safety disclaimer.
Improve, Shorten, Translate, rewrite for an HCP or patient audience. Output is quality-checked before it lands; some actions show a proposal you accept or reject.
Name the expert inline in the page and the Chief of Staff steps aside.
@regulatory_strategist inline in the page body — no recommendation card.
Two is the edge. Add a third for a panel, or pick the stronger fit.
Attributed contributions; the synthesis preserves dissent.
Escalates to Deliver, with gates.
An inserted answer arrives AI-drafted; the moment you touch it, it is reclassified as human-edited. Agents are versioned like colleagues are named.
When the Chief of Staff or an expert draws on knowledge or a connector, claims are cited and what was fetched is saved under your work — your audit trail, reusable later.
Standard content exports faithfully; shared pages render read-only with full fidelity.
Work moves between colleagues the way it moves inside a real team.
Never out to a tool and back. A tool can join a workflow — only a workforce can share the work.
Skeptics make the best reviewers — and reviewers are who we built this for. Put the first record on your own work.