Missed Opportunity Analysis // G2 Intent + CRM

Identifies accounts showing buyer intent for competitor products that are not yet in the CRM, so sales teams can prioritize outreach on prospects most likely to be in-market. It removes the manual work of cross-referencing intent signals against CRM records and filtering for ICP fit.

1. Query the G2 MCP Server to retrieve all accounts with buyer intent signals in the last N days, capturing company name, domain, product, signal type, and intent score.
2. Filter the results to retain only accounts showing intent against competitor products (not your own product), flagging any accounts with both competitor and your own product intent.
3. Apply firmographic filters (industry, employee count, revenue, geography) to narrow to ICP-confirmed accounts; flag accounts with missing firmographic data separately.
4. Query your CRM by domain to match each account; sort accounts not found in CRM to the top of your output. Output: a table of high-intent prospects with confirmed ICP fit that are missing from your CRM, ranked for outreach priority.

Sales teams focus outreach on in-market prospects already showing competitor intent, without manual cross-referencing of intent data and CRM records.

• G2 MCP Server connector: required to retrieve buyer intent signals, signal types, and intent volume scores for accounts in the last N days.
• CRM MCP connector for your CRM system (e.g., Salesforce, HubSpot): required to query existing account records by domain and retrieve owner and stage information.
• Prompt template used by this workflow:

"You are running a G2 buyer intent to CRM gap analysis. Work in four discrete steps. Do not skip a step or merge steps to save calls, each filter is a hard gate, not a scoring weight.

Step 1, Pull raw intent Using the G2 MCP Server, retrieve all accounts showing buyer intent in the last {INTENT_WINDOW_DAYS} days. Capture for each account: company name, domain, product(s) with intent signal, signal type (e.g. category views, comparison views, alternative page visits), and intent volume/score as returned.

Do not aggregate or de-duplicate yet, keep raw output.

Step 2, Filter to competitor intent From the Step 1 output, retain only accounts whose intent signal is against one or more products in {COMPETITOR_LIST}. Discard signal against your own product or unrelated categories. If an account has intent on both a competitor and your own product, keep it and flag it, that's a switch signal, not noise.

Step 3, Filter to ICP Apply firmographic filters to the Step 2 list:

Industry in {ICP_INDUSTRY} Employee count within {ICP_EMPLOYEE_RANGE} Revenue within {ICP_REVENUE_RANGE} (if set) Geography in {ICP_GEOGRAPHY}

Use whatever firmographic fields the G2 MCP returns natively. If firmographic data is missing or unresolvable for an account, do not guess the fit, list it separately as 'ICP unconfirmed' rather than including or excluding it silently.

Step 4, Match against CRM, isolate gaps Query {CRM_SYSTEM} ({CRM_OBJECT}) via its MCP connector for existing records matching each Step 3 account by {CRM_MATCH_FIELD} (domain match, not name match, name matching produces false positives on common company names). For each account:

If a matching CRM record exists, mark 'In CRM' and note owner/stage if present. If no matching record exists, mark 'Not in CRM', this is the target output.

Output Produce a table with columns: Company | Domain | Competitor(s) with intent | Intent signal type/volume | ICP fit (Confirmed / Unconfirmed) | CRM status (In CRM, owner/stage / Not in CRM) | Recommended next action

Sort 'Not in CRM' + 'ICP Confirmed' accounts to the top, that's the actionable list. Do not editorialize the intent data or infer purchase likelihood beyond what the signal type indicates."

Relies on G2 and CRM data quality; missing or inaccurate firmographic data prevents ICP confirmation. Domain matching avoids false positives on common company names, but accounts with unknown or non-standard domains may be unresolved. Intent signal filtering treats each filter as a hard gate, so accounts at the ICP boundary or with incomplete data are flagged separately and require manual review rather than automatic inclusion.