---
name: ICP Qualification Framework
slug: icp-qualification
description: How agents score lead fit (A/B/C/disqualify) consistently across inbound, MQL, outbound, and ABM.
category: qualification
---

# ICP Qualification Framework

Every agent in the hub uses the same 4-tier scoring so handoffs are predictable.

## The four tiers

- **A — Strong fit, high intent.** Send to a rep now. SLA: same day.
- **B — Strong fit, low intent OR medium fit, high intent.** Nurture sequence + watch
  for second signal.
- **C — Borderline.** Add to long-term nurture only. No outbound spend.
- **Disqualify** — Wrong company, wrong region, blocked competitor, free-tier
  consumer. Mark in CRM, never re-touch.

## Fit dimensions

Score each on a 0/1/2 (no/maybe/yes), then sum:

1. **Industry / segment** — is the company in a segment we explicitly sell to?
2. **Size / stage** — headcount, funding, revenue band match?
3. **Geography** — in a region we operate in / can support?
4. **Title / persona** — is the person in a buying role (not student, not job-seeker)?
5. **Tech / data signal** — do they use a system / product that suggests fit?

Sum:
- 8–10 → A-fit
- 5–7 → B-fit
- 3–4 → C-fit
- 0–2 → Disqualify

## Intent dimensions

Score signals (most recent wins):

- **High**: pricing page visit, demo request, "how do I buy" question, RFP, signed
  up + active, talked to a competitor's rep.
- **Medium**: downloaded gated content, attended webinar, opened+clicked sequence
  twice, visited site 3+ times in a week.
- **Low**: opened one email, followed on social, nothing recent.

A-fit + High intent = A-tier lead.
A-fit + Low intent = B-tier (warm nurture).
B-fit + High intent = B-tier (watch for second signal before SDR spend).

## Hard disqualifiers (override everything)

- Competitor employee.
- Student / job-seeker / academic researcher (unless we sell to that segment).
- Region we don't sell to.
- Free email domain on a "request enterprise demo" form (almost always low signal).
- Already a customer of our paid plan (route to CS, not sales).

## How to log

Every agent that scores must write back to CRM:

- `lead_tier`: A | B | C | DQ
- `fit_reasoning`: 1 sentence
- `intent_reasoning`: 1 sentence + the specific signal
- `recommended_next_step`: route_to_rep | nurture | disqualify | wait_for_signal
- `agent_run_id`: which agent run produced this score (for audit)

## Anti-patterns

- Don't conflate fit and intent. A-fit with no intent is *not* an A-lead — it's a
  great account to nurture.
- Don't score off enrichment data alone. Read the actual form note / source page.
- Don't keep re-scoring the same lead every week. Score on event, not on cron.