The problem this solves
At the end of this mission every company in your HubSpot portal carries two new properties: an ICP fit score from 0 to 100 and the top reason behind it. Reps sort their views by the score, workflows branch on it, and a Google Sheet shows the whole book of business ranked with a chart of how the scores spread. That chart matters more than it sounds: a portal where most companies score under 30 is telling you something about the ICP or the database, and a picture makes it impossible to miss.
Portals accumulate companies from every source for years - imports, form fills, integrations, reps adding accounts by hand - and the score that would tell them apart usually depends on fields that were never filled. Scoring against live firmographics and technographics instead of the portal's own properties means a company with an empty industry field is evaluated on what the company actually is, with the evidence that produced the score stored beside it.
Writing scores back is the step that makes this permanent, and the step that deserves care. A bulk property update across thousands of records has to be reviewable and reversible, which is why the mission stages the write and keeps the before-and-after values.
How the mission runs
- Pull the company records and resolve domains. The Lead Verification Agent reads every company from HubSpot with its domain, name, industry, employee count, lifecycle stage and owner. Companies without a domain are resolved by name and location through company search, and any that still cannot be resolved are grouped separately so they can be scored on what the portal knows or skipped.
- Refresh the evidence for every company. Firmographics and technographics are pulled for each resolved domain with waterfall enrichment, so each field is filled by the first source that answers and only answered lookups are charged. The evidence is kept per company so the score can be explained now and refreshed later.
- Score 0-100 against the ICP. Each company is scored on the rubric proposed from your ICP - the agent shows the dimensions and weights before scoring - and receives a score, a full reasons list and a single top reason. The top reason is written in plain words a rep can read in the record view, such as headcount and industry match with no technology overlap.
- Stage the write-back to custom properties. The agent proposes two company properties, ICP fit score as a number and ICP fit reason as text, creating them if they do not exist, and stages the update for every company as one batch. The batch shows how many records will change and the previous values where a score already existed, and waits for your approval before any property is written.
- Build the ranked book of business in Google Sheets. A sheet lists every company ranked by score with owner, lifecycle stage, top reason and the evidence columns. A distribution chart of the scores, rendered by the platform, sits on a summary tab with the count per band, and the sheet link plus the credit total closes the mission.
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
Two custom company properties in HubSpot, ICP fit score and ICP fit reason, populated on every scored company after your approval of the staged batch. A Google Sheet with the ranked book of business - every company with owner, stage, score, top reason and evidence - and a summary tab with the score distribution charted and counts per band. A change log of previous and new property values so the write can be reviewed or reversed, and the full audit trail of enrichment calls with their credits.
Make it yours
- Limit the scope to companies owned by one team or in one lifecycle stage, so a first run scores the open pipeline before the whole portal.
- Write a tier property alongside the number, so HubSpot workflows can branch on A, B and C without arithmetic.
- Run it monthly on a schedule and ask for a movement column, so the sheet shows which companies rose or fell since the last run.
- Replace HubSpot with Pipedrive or Attio; custom fields take the place of properties and the rest of the mission is unchanged.
Frequently asked questions
Will this overwrite scores my team set by hand?
Only after you see them. The staged batch lists every company where a value already exists, with the old and new score side by side, and you can exclude those rows or keep the manual score in a separate property.
How long does a full portal take?
It depends on the number of companies and how many need enrichment. The agent reports the record count and the credit estimate before it starts, and works through the portal in batches so partial progress is visible in the mission log.
Can HubSpot workflows use the new properties?
Yes. Once the properties exist and are populated, they behave like any other company property: list filters, workflow enrollment triggers and view columns all work with them.