Deal Brain Concept
Fig. 1 · Proposal, drawn from history

Quote what the work will actually take.

Deal Brain reads a new brief, pulls the comparable jobs you have already delivered with their real hours and margins, scores the tender's complexity, and drafts the first offer. From your history, not a blank page.

New brief · just landed Complexity 7.2 / 10
Operator manual set, six languages, 340 pages, CE technical file
DE source Q4 delivery Illustrations reworked
Comparables · 3 of 212 · illustrative Quoted / actual
Operator and service manuals, five languages
2024 · match 92 · margin 27%
180 h / 214 h
Won
Spare-parts catalogue, six languages
2023 · match 81 · margin 34%
240 h / 236 h
Won
Machine documentation with CE file, four languages
2025 · match 74 · priced 18% under comparables
150 h / n/a
Lost
Assumptions carried: 6 · flagged: 2 Draft first offer
Fig. 2 · The blank page problem

Work is priced before its effort is understood.

Note 01

A quote starts from a blank page. There is no coded library of past proposals, assumptions and outcomes to start from.

Note 02

Proposals live in individual inboxes and drives. Each seller re-derives complexity, wording and margin logic from memory.

Note 03

What was quoted and what delivery later discovers the work required drift apart. Nobody sees the gap until the job is done.

"No institutional memory today. Every proposal relies on individual context."

Leadership ask, discovery interviews
Fig. 3 · Process, four stations

A brief lands. History answers first.

Station 01
Retrieve the comparables

Reads the brief, pulls the nearest delivered jobs with what actually happened: scope, assumptions, hours quoted against hours spent, margin, won or lost.

Station 02
Score the complexity

Tender complexity is scored against that history, so a hard job looks hard before it is priced, not after it is delivered.

Station 03
Draft the first offer

A consistent first draft from a coded proposal library, in house wording, every assumption written down. Commercial approval stays with people.

Station 04
Route the genuinely new

No comparable found: it goes to a delivery-team review instead of a guess at scope. Their answer becomes the next comparable.

It compounds. Every quote and its outcome feeds back in as comparable-job data. The thing the business lacks at the point of commitment today is what Deal Brain accumulates with every use.

Fig. 4 · Bill of asks, nine items

Nine asks from seven directions, one drawing.

Every idea logged in discovery that points at this tool, and the station on the drawing where it lands.

Ref Ask What it asked for Source Station
KR-05
Proposal and quotation library

Store coded offers, assumptions and outcomes so comparable proposals can be found and reused.

Founder register 01 Retrieve
CA-01
Quote support from history

Find comparable work, assumptions, scope and outcomes before a new quotation is prepared.

Founder register 01 Retrieve
CA-02
Generate consistent offers

A consistent first draft from a controlled proposal library, with commercial approval kept with staff.

Founder register 03 Draft
N-11
Speed to proposal

Speed up proposal generation and aggressively win more clients.

Founder atlas 03 Draft
A-06
Turn tribal knowledge into a tool

Capture the judgement in a few senior heads: surface similar past jobs when a request lands, draft quotes and feasibility assessments from history, score tender complexity.

Opportunity document 01 + 02
W-01
Deal Brain

An agentic proposal builder: a new hire generates a 90% proposal from historical deals. Assumes access to proposal history.

Comments, wild future 03 Draft
U-05
Quote-to-proposal agent

The same agent, already promised in the partner brief's wave one. Deal Brain seeds it rather than competing with it.

Use-case inventory, partner briefs 01 to 04
B-01
Proposal builder as the lighthouse

No institutional memory today. A new hire should generate a 90% proposal from historical data; wins on pricing accuracy and new business.

Leadership asks Compounds
S-01
Proposal drafting from tech specs

Draft proposals straight from the technical specification that arrives with the request.

Sales focus group 02 + 03
Fig. 5 · Survey of the space, September 2026

Each half exists somewhere. Nobody joins them.

Response platforms draft from a content library. Proposal builders template and track. Language-industry systems price by word count. Estimating from actuals exists in construction and professional-services software. None of them reads what a documentation job actually cost you.

Capability
Response platformsLoopio, Responsive, Tribble, Arphie
Proposal buildersPandaDoc, Qwilr, Proposify
LSP systemsPlunet, XTRF
Deal Brainbuilt on your outcomes
Drafts from a content library
Templates, tracking, approval workflow
Prices from rate cards and word counts
Retrieves comparable delivered jobs with actual hours and margin
Scores tender complexity against history
Learns from your quoted-versus-actual outcomes
Core Partial Not offered
Nearest analogues

Loopio for library-driven first drafts, Qwilr for structured and trackable proposals, PandaDoc for content blocks with approval. Each covers the writing half of the job.

Where the pattern lives

Estimating from a completed-job library is normal in construction estimating and in professional-services platforms such as Deltek and Kantata. Not in documentation or translation services.

Already in the building

Plunet generates Global Content offers from rate cards today. Deal Brain reads its outcomes; it does not replace it. AI-native response tools start around $30k a year and still draft from text.

Fig. 6 · Objections, annotated
Does it set the price?

No. It removes the blank page and gives every quote a documented, checkable starting point. Commercial judgement stays with the people who own the margin.

What happens when a senior seller leaves?

The judgement that lives in a few senior heads becomes retrievable. The target set in discovery: a new hire drafts a 90% proposal from historical deals in their first week.

Why not buy one of the tools above?

They draft from a content library. Deal Brain drafts from outcomes: what a job actually cost against what you quoted. That data exists nowhere else, and no vendor can ship it.

Fig. 7 · Convergence

Asked for nine times, from seven directions.

Sales, Engineering, Experience and Global Content each raised it independently. No other idea in the discovery converged this hard.

9 ideas pointing at the same tool
7 independent sources, from founder register to sales focus group
4.65 opportunity score, the highest in the atlas
Source: discovery opportunity atlas, September 2026
Fig. 8 · Build schedule

Ten proposals. Five days.

Nothing to buy and nothing to migrate. The first version runs on a folder of past offers and the outcomes you already know.

Request early access
Days 1 to 2

Build the schema. Ingest ten real proposals with their assumptions and outcomes, or a synthetic set if the real ones are still in inboxes.

Days 3 to 5

Add retrieval over the library and the first-draft screen. Score complexity on the ten and compare it with what delivery remembers.

Needs

Proposal history from Sales. Ten offers is enough to start; two hundred is where it gets interesting.

Deal Brain
Sheet 1 of 1 · Concept page · internal build proposal · not a product