Target Account Scanner

Adam's take

If you only build one agent this year, build this one. Every downstream agent inherits the quality of the universe it creates, and most teams discover their 'ICP' was three opinions in a trench coat the moment they have to write the config.

Required tools

Optional tools

Expected value

Replaces 10-20 hrs/month of manual account research; consistent ICP coverage

Deployable workflow pack

The full pack includes the agent spec, deployment guide, architecture diagrams, n8n workflow skeleton, Clay template, HubSpot properties, JSON schemas, OpenAPI spec, cost and ROI calculators, demo data, QA checklist, and governance notes.

How to deploy this agent

Most teams define an ICP once, in a slide, and never operationalize it. This agent turns the ICP definition into a versioned, running scan against firmographic, technographic, and intent data, and writes the result back to HubSpot as agent-owned properties.

Core signals

Input schema

{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "$id": "https://deploytheagent.com/schemas/agent-001/input.schema.json",
  "title": "Target Account Scanner — Input",
  "description": "One candidate account record entering the scoring step. Enrichment fields may be null (scored neutral or zero per workflow logic); required identity and CRM-gate fields may not.",
  "type": "object",
  "required": ["domain", "company_name", "crm_status"],
  "additionalProperties": false,
  "properties": {
    "domain": {
      "type": "string",
      "pattern": "^[a-z0-9]([a-z0-9-]*[a-z0-9])?(\\.[a-z0-9]([a-z0-9-]*[a-z0-9])?)+$",
      "description": "Normalized root domain, lowercase, no protocol or www"
    },
    "company_name": { "type": "string", "minLength": 1 },
    "industry": { "type": ["string", "null"] },
    "employee_count": { "type": ["integer", "null"], "minimum": 0 },
    "revenue_usd": { "type": ["integer", "null"], "minimum": 0 },
    "hq_country": { "type": ["string", "null"], "description": "ISO 3166-1 alpha-2" },
    "hq_region": { "type": ["string", "null"] },
    "tech_stack": {
      "type": "array",
      "items": { "type": "string" },
      "default": []
    },
    "hiring": {
      "type": "object",
      "additionalProperties": false,
      "properties": {
        "open_roles_target_function": { "type": ["integer", "null"], "minimum": 0 },
        "functions": { "type": "array", "items": { "type": "string" } }
      }
    },
    "funding": {
      "type": "object",
      "additionalProperties": false,
      "properties": {
        "last_round": { "type": ["string", "null"] },
        "last_round_date": { "type": ["string", "null"], "format": "date" },
        "last_round_amount_usd": { "type": ["integer", "null"], "minimum": 0 }
      }
    },
    "intent": {
      "type": "object",
      "additionalProperties": false,
      "properties": {
        "level": { "type": ["string", "null"], "enum": ["surging", "moderate", "none", null] },
        "topics": { "type": "array", "items": { "type": "string" } },
        "provider": { "type": ["string", "null"] }
      }
    },
    "crm_status": {
      "type": "object",
      "required": ["lifecycle_stage", "has_open_deal"],
      "additionalProperties": false,
      "properties": {
        "lifecycle_stage": { "type": ["string", "null"] },
        "has_open_deal": { "type": "boolean" },
        "is_customer": { "type": "boolean", "default": false }
      }
    },
    "source": {
      "type": "string",
      "description": "Where the candidate came from: apollo, import, web_visitor, manual"
    }
  }
}

Output schema

{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "$id": "https://deploytheagent.com/schemas/agent-001/output.schema.json",
  "title": "Target Account Scanner — Output",
  "description": "One scored account record leaving the workflow. Written to HubSpot agent-owned properties and included in the Slack digest.",
  "type": "object",
  "required": [
    "domain",
    "icp_match_score",
    "icp_match_band",
    "icp_match_reasons",
    "icp_scan_date",
    "icp_model_version",
    "gate_result",
    "score_breakdown"
  ],
  "additionalProperties": false,
  "properties": {
    "domain": { "type": "string" },
    "icp_match_score": { "type": "integer", "minimum": 0, "maximum": 100 },
    "icp_match_band": { "type": "string", "enum": ["Target", "Monitor", "Exclude"] },
    "icp_match_reasons": {
      "type": "array",
      "items": { "type": "string" },
      "minItems": 1,
      "description": "One line per scoring dimension, with source attribution"
    },
    "icp_scan_date": { "type": "string", "format": "date" },
    "icp_model_version": {
      "type": "string",
      "description": "Config version plus prompt version when LLM reasons are enabled, e.g. icp-config-1.0.0/reasons-off"
    },
    "gate_result": {
      "type": "string",
      "enum": ["written", "skipped_customer", "skipped_open_deal", "skipped_excluded_domain"],
      "description": "Whether the record was written to CRM or suppressed, and why"
    },
    "score_breakdown": {
      "type": "object",
      "required": ["firmographic", "technographic", "signals"],
      "additionalProperties": false,
      "properties": {
        "firmographic": { "type": "integer", "minimum": 0, "maximum": 55 },
        "technographic": { "type": "integer", "minimum": 0, "maximum": 15 },
        "signals": { "type": "integer", "minimum": 0, "maximum": 30 }
      }
    },
    "confidence": {
      "type": ["number", "null"],
      "minimum": 0,
      "maximum": 1,
      "description": "Share of scoring fields that were non-null; low confidence flags enrichment gaps, not bad fit"
    }
  }
}

Scoring / classification logic

Accounts scoring 70+ land in the Target band and get written to HubSpot. 50-69 is Monitor. Below 50 is Exclude. Every scoring reason is logged per account with source attribution, and the config version is stored alongside the score so a team can trace exactly which weights produced a given result.

HubSpot CRM properties

n8n workflow skeleton

Weekly ICP scan: pull candidates from Clay, score against ICP config, gate on CRM status, write agent-owned HubSpot properties, post Slack digest. Configure every node marked CONFIGURE in its notes.

1. Weekly scan trigger
   CONFIGURE: default Monday 06:00. First universe build: run manually.
2. Load ICP config
   CONFIGURE: paste your versioned icp-config JSON here or fetch from repo raw URL. Version string flows to icp_model_version.
3. Fetch candidates from Clay
   CONFIGURE: Clay table webhook/HTTP API URL and {{CLAY_API_KEY}}. Expects rows matching clay-template.csv headers.
4. Validate against input schema
   Drops rows failing schemas/input.schema.json shape; logs drops. Do not silence the log.
5. Score ICP match
   Implements the scoring table from agent-spec.md. Weights come from config, not code edits.
6. CRM status gate
   Suppression first: customers and open-deal accounts are never written. Gate failures block the run rather than passing records through.
7. Write agent-owned properties
   CONFIGURE: {{HUBSPOT_CREDENTIAL}}. Writes ONLY the five icp_* properties from hubspot/properties.json. Batch 100 per call. Only gate_result=written records reach this node (add an IF node filter on import if your n8n version needs it explicit).
8. Post Target-band digest
   CONFIGURE: {{SLACK_CREDENTIAL}} and channel. Digest lists new Target-band accounts with score, top reasons, and a review link. Humans approve before any outbound use.
9. Write run log
   Audit trail: run date, config version, counts per band, per-field fill rates, gate skips. Alert if Target band >25% of universe or no run in 8 days (pair with a separate monitor workflow).

Agent-driven Clay integration (optional)

The default deployment path pulls Clay data through a static export URL, shown above. Teams already running Clay through its official CLI/MCP (clay-run/agent-plugins) can use this agent-driven path instead.

Cost calculator

ROI calculator

Governance notes

• Suppression runs before scoring reaches CRM: accounts already tagged customer or with an open deal are never written, regardless of score (gate_result: skipped_customer / skipped_open_deal).
• Writes land only in agent-owned HubSpot properties — never an existing lead-scoring or lifecycle-stage field.
• Every scored record carries a model/config version and a scan date for auditability.
• The workflow does not create CRM contacts. Contact-level sourcing is a separate, reviewed process (Agent #005).

Failure modes

• Missing firmographic data: nulls are scored neutral where the config defines a neutral default (e.g. unknown revenue), not zero — this avoids penalizing gaps in enrichment as if they were poor fit.
• Stale or missing intent: no intent data scores as "none," not as a false positive.
• Duplicate accounts: rows are deduplicated by domain before scoring.
• Runaway Target band: if more than roughly a quarter of the scanned universe lands in Target, that signals the config is too loose and should be retuned before the list is trusted for outbound.

QA checklist

FAQ

Related agents