A literary agency where reading a submission and writing it up was two separate days of work

An agent reads a long document end to end: a manuscript, a rights contract, an offer, a co-agency agreement. The reading is the job and nobody wants to shortcut it. But once it is read, someone still has to sit down and retype the whole thing into a synthesis sheet: the premise, the terms, the comparable titles, the open questions, the recommendation. That write-up, not the reading, was the step that ran into a second day.

Meanwhile every agent worked their own pole. Each one held their own authors, their own submissions, their own deal terms in their own files, and the agency's records lived across a stack of spreadsheets plus a shared drive. As the roster grew, finding the current version of a deal, knowing who was on which submission, and getting a new coordinator up to speed all got slower. On top of it, the team estimated that a real share of the inbox was email that never needed to be sent, threads chasing a status that already existed somewhere nobody could see.

We started with the foundation the team kept circling back to: one platform as the single source of truth. Every author and every submission in one place, a structured intake so the same details are entered once at the right spot, the people attached to each file, role-based access, notes kept by role for the agent, the coordinator, and the rights manager, and a dashboard showing the status of every submission and deal at a glance.

On top of it sits the flagship: a reading-to-synthesis tool. You drop in the transcribed read or the document itself, click structure, and the summary sheet appears, premise, terms, comparable titles, open questions, recommendation, every field editable, grouped the way an agent actually thinks, and exportable in seconds. The extraction is AI-driven with a plain rules-based fallback behind it, so it produces a usable sheet even when the source is messy and never simply fails. Because the status of every file now lives in one view, the reflexive status-chasing emails mostly stopped sending themselves.

The write-up that used to spill into a second day now takes minutes to draft and a short while to refine. Agents read the same volume of submissions, but the part a person spends turning a finished read into a clean sheet dropped by about eighty percent, and the work left for a person is the judgment, deciding and adjusting, not retyping.

The agency also stopped hunting across a stack of spreadsheets for the current terms. There is one source now, the status of every submission and deal is visible in one view, and a new coordinator finds their way around in days instead of weeks. The email churn fell with it, because most of those threads existed only to ask a question the dashboard now answers.