Account-based marketing was always more aspirational than practical for most B2B teams because the per-account labor was prohibitive. AI changes that math. Real ABM at meaningful scale is now possible for mid-sized B2B teams. Here is the practical workflow.
Traditional ABM was: pick 100 dream accounts, research each deeply, create custom content for each, run targeted ads, coordinate sales and marketing handoffs. The per-account labor was so high that most teams could only do it for 20-30 accounts.
AI compresses the research and personalization labor by 80-90%, which means real ABM at 100-500 accounts is now feasible for a 5-15 person GTM team. This changes the math of who should do ABM.
1. Account selection with AI scoring. Load your operational ICP into a Claude Project. Run your prospect list through it. Get back ranked accounts with reasoning. See how to use AI to research prospects.
2. Tier the accounts into 3 cohorts. Tier 1 (10-20 dream accounts, deep custom work), Tier 2 (50-150 accounts, semi-custom), Tier 3 (200-500 accounts, automated personalization).
3. Generate account briefs at scale. Claude generates a structured account brief per account using web data + your ICP framework. 5-10 minutes per account vs. 90 minutes manually.
4. Build per-account messaging variants. For Tier 1: custom landing page + sequence. For Tier 2: persona-specific sequence with account-level personalization. For Tier 3: ICP-specific sequence with light personalization.
5. Coordinate sales/marketing on shared briefs. Sales and marketing both reference the same Claude-generated account brief. Eliminates the "what is sales pitching this account on?" disconnect.
Generate an ABM account brief for [TARGET COMPANY] using the URL [URL]. Our ICP fields (from knowledge base) are: [auto-loaded] Our unique selling angle: [SHORT VERSION] Return: 1. Account fit score (1-10) with reasoning against our ICP 2. Recent timing signals (last 90 days: funding, hires, product launches, churn signals, regulatory changes) 3. Likely buying committee (named where possible, role-typed otherwise) 4. Top 3 angles for outreach based on the specific account, not generic value-prop 5. Account-specific risks (why this might NOT close) 6. Recommended Tier (1, 2, 3) and recommended motion intensity Flag any data we should not trust without verification.
Required: Claude Team (the Project layer), Perplexity Pro (for real-time data), your existing CRM, your existing sales engagement tool.
Optional but powerful: 6sense or Demandbase for intent data, Apollo or Cognism for contact enrichment, Mutiny or Pathfactory for custom landing pages.
Most teams over-buy ABM tooling. The core capability is research + personalization at scale. Claude + Perplexity + your existing CRM covers that. Add specialty ABM platforms only when you have a specific gap they fill.