Market and ICP Reset
Define market choice and ICP criteria so AI can read, score, and route accounts.
Key takeaways
- A good ICP is operational data, not narrative: fields, scores, source links, exclusion rules, and routing logic that AI can read and repeat.
- Define ICP in four layers (firmographic, technographic, trigger, pain/outcome) and add a proof-difficulty axis to segmentation.
- Keep fit, intent, readiness, and expansion as separate scores, then combine them with weights (0.30F + 0.30I + 0.25R + 0.15E).
- Route by score: 4.0+ goes to AE within 24 to 48 hours, 3.2 to 3.9 gets data verification, 3.1 or lower goes to nurture or exclusion.
- Make exclusion rules explicit and recalibrate weights monthly using meetings, POCs, wins, losses, expansion, and churn.
The first step in AI GTM is deciding who to sell to with more precision. AI can gather account information at scale, but without explicit criteria it only creates plausible-looking lists. A good ICP is understandable to sales, recordable in CRM, and repeatably scoreable by AI.
Define ICP in Four Layers
| Layer | Question | Example criteria |
|---|---|---|
| Firmographic | What kind of company is it? | Industry, size, region, revenue, growth stage |
| Technographic | What systems do they use? | CRM, warehouse, cloud, collaboration tools |
| Trigger | Why now? | Funding, hiring, regulation, cost pressure, system change |
| Pain/Outcome | What outcome do they need? | Forecast accuracy, cost reduction, productivity, risk reduction |
AI-ready ICP
AI-ready ICP is operational data: fields, scores, source links, exclusion rules, and routing logic.
Segment by Proof Difficulty
Traditional segmentation uses company size and industry. AI GTM adds one more axis: how hard it is to prove value.
| Segment | Buying power | Proof difficulty | GTM motion |
|---|---|---|---|
| Clear ROI SMB | Low to medium | Low | Self-serve, templates, automated onboarding |
| Mid-market Workflow | Medium | Medium | Sales assist, product-signal routing |
| Enterprise Transformation | High | High | FDE, security review, POC, executive sponsor |
| Regulated Buyer | High | Very high | Compliance, audit, data separation, long procurement |
Separate Fit From "Now"
Many teams mix fit score and intent score. That hides the difference between a great account that is not buying now and a noisy account that is active but structurally poor fit.
| Score | Meaning | Data examples |
|---|---|---|
| Fit score | Is the account structurally right? | Industry, size, stack, use case |
| Intent score | Is the account trying to solve this now? | Search, page visit, event, hiring, tech change |
| Readiness score | Can the account adopt? | Data quality, budget, owner, security requirements |
| Expansion score | Can it expand? | Teams, workflows, usage, subsidiaries |
| Result | Routing |
|---|---|
| 4.0+ | AE or sales assist within 24 to 48 hours |
| 3.2 to 3.9 | SDR or RevOps verifies data and trigger before nurture or discovery |
| 3.1 or lower | Automated nurture, long-term watch, or exclusion check |
Make Exclusions Explicit
AI is good at finding positive signals. Without exclusion rules it wastes seller time.
- Accounts that cannot accept the deployment model for security reasons
- Accounts whose budget does not match pricing structure
- Accounts where the decision owner does not own the category
- Accounts that cannot provide data needed for a successful POC
- Accounts whose custom needs would pull the roadmap too far
ICP Operating Loop
### 1. Review the last 10 win/loss accounts. Look first at buying reason and adoption blockers, not deal size.
### 2. Convert common signals into fields. Keep only signals that can be recorded in CRM, product logs, web behavior, or external events.
### 3. Separate scoring from routing. Score is priority; routing decides owner, playbook, and SLA.
### 4. Recalibrate monthly. Adjust weights using meetings, POCs, wins, losses, expansion, and churn.
ICP Canvas
| Item | Working answer |
|---|---|
| Primary segment | First industry, size, and region to target |
| Urgent trigger | Event that starts buying consideration |
| Economic buyer | Role that owns budget and priority |
| Technical blocker | System or security condition that can block adoption |
| Proof metric | Metric the POC must prove |
| Exclusion rule | Condition that removes an account from sales motion |
Operating Checklist
- ICP is represented as fields and scores, not just narrative.
- Fit, intent, readiness, and expansion scores are separate.
- Exclusion rules affect routing.
- AI-generated account lists include source links and signals.
- Win/loss review updates ICP weights.