Turn a new assigned account into a 15-minute research brief

Start by exporting newly assigned accounts from HubSpot, folk, Airtable, or your territory spreadsheet into one working table. Create fields for account name, domain, owner, territory, segment, industry, employee_count, region, current_status, existing_contacts, last_activity_date, source, and research status. Deduplicate by domain, normalize company names, and remove students, agencies, vendors, competitors, existing customers, and accounts with an active owner conflict. Add a notes field for missing data so the rep does not confuse an incomplete record with a weak account. QA check: every account must have a domain, owner, segment, and keep/remove decision before research begins.

Create a clean research-ready account table from this account list.

Raw account list:
{{raw_account_list}}

Required fields:
- account_name
- domain
- owner
- territory
- segment
- industry
- employee_count
- region
- current_status
- existing_contacts
- last_activity_date
- source
- research_status

Output a cleaned table and flag duplicates, missing domains, unclear owners, obvious non-fits, and accounts that need manual review.

Use Claude to score each account against your ICP before anyone spends time reading websites and LinkedIn profiles. Use a simple 1-5 fit score with reasons tied to firmographic, technographic, geographic, industry, and trigger criteria. Add fields for icp_score, fit_reason, exclusion_reason, missing_information, and research_priority. Accounts with low fit or missing critical data should go to a low-priority queue instead of receiving a full brief. QA check: review 10 scored accounts with a manager to confirm the scoring matches your real territory strategy.

Score these accounts against our ICP.

ICP summary:
{{icp_summary}}

Accounts:
{{clean_account_table}}

For each account, return:
1. ICP fit score from 1-5
2. Fit reason
3. Disqualification or exclusion reason if any
4. Missing information
5. Recommended research priority: high, medium, low, or skip
6. What a rep should verify manually

Use conservative scoring. Do not assume fit when key data is missing.