The problem this solves
The quarterly planning question - where should pipeline come from - is usually answered with last quarter's segments and a feeling about which ones felt easier. The evidence for a better answer exists: the website says what the product does and for whom, the closed-won accounts say who actually bought, and market data says how many companies of each kind exist and what they run. Put together, they rank the segments by fit with a reason for each rank.
The Company Intelligence Agent does the putting together. It reads the site, enriches the closed-won accounts, derives the attributes that winning customers share, defines six to eight candidate segments with sizes from live company data, scores each for fit against the winning pattern, charts the ranking and writes a short deck that a sales leader can carry into the planning meeting. Every score shows its evidence, so the deck can be argued with rather than merely accepted.
How the mission runs
- Read the website and the won accounts. The agent crawls <mysite.com> for the offer, the price point and the functions addressed, then reads HubSpot for closed-won accounts and enriches them with industry, headcount, region, stack and funding stage so the winning pattern can be derived from data rather than memory.
- Derive the winning pattern. Across the won accounts the agent finds the attributes that recur - the industries, the size band, the technologies, the stage - and weights each by how consistently it appears, producing a pattern that is shown before any segment is scored.
- Define six to eight segments with sizes. Candidate segments are defined by industry, size and technology pattern, and company search sizes each one from live firmographic and technographic data, so the deck says how many companies each segment holds as well as how well it fits.
- Score and chart the segments. Each segment is scored 0-100 against the winning pattern with the evidence per score, and the chart capability renders the ranking as a bar chart with segment size as a second axis, so fit and opportunity are visible together.
- Write the deck. A short deck in Google Docs is written - one page of method, one chart of the ranking, one page per top segment with its score, size, evidence and the message it would respond to, and a closing page recommending where next quarter's pipeline should come from.
The prompt
This is the exact objective the agent receives. Swap the obvious placeholders for your own domain, segment or channel and run it as-is from the console, Slack, or the API.
What comes back
A short deck in Google Docs ranking six to eight market segments by fit against your closed-won pattern, with a chart of fit against segment size, one page per leading segment showing the score, the evidence, the number of companies and the message that fits, and a recommendation for next quarter's pipeline focus. The segment definitions are saved so lists can be built from the winners immediately.
Make it yours
- Include closed-lost accounts so the pattern reflects what fails as well as what wins.
- Build a 50-company list from the top segment in the same mission as proof of supply.
- Re-run quarterly and add a page showing how the ranking moved.
- Deliver the deck to Notion or as a PDF in Drive for the board pack.
Frequently asked questions
How many closed-won accounts are needed?
A pattern emerges from a dozen; more is better. With very few, the agent leans on the website and says so, and the scores carry wider confidence bands.
Can I add my own segment hypotheses?
Yes. Name them and they are sized and scored alongside the agent's own, which is often the most useful comparison in the deck.
Is the segment size an estimate?
It is a count from live company data under the segment's definition, stated with the definition so the count can be checked.