Don't start with messaging. Build a hyper-specific list based on observable public data that signals a clear pain point. This data-driven list itself becomes the core of a highly relevant message, moving beyond generic persona-based outreach and hollow personalization.

Traditional ICP scores reflect who *you* want to sell to (e.g., wallet size), which is useless for reps. Instead, sort your entire market based on the quantifiable size of their pain (e.g., projected fines). This gives reps a clear, actionable, and customer-centric reason for outreach.

Go beyond simple prompts. Gather raw data—comments from your social media, competitor book reviews, and podcast feedback—and feed it all into ChatGPT. Then, ask it to synthesize this data into a detailed avatar guide, identify market gaps, and suggest opportunities for your offer.

Instead of manually sifting through overwhelming survey responses, input the raw data into an AI model. You can prompt it to identify distinct customer segments and generate detailed avatars—complete with pain points and desires—for each of your specific offers.

Executive teams often create an ICP based on a 'wishlist' of big logos. The most accurate ICP is actually found by analyzing your first-party CRM data. Examining patterns across both close-won and close-lost deals reveals surprising truths about which customer segments are actually the best fit for your solution.

Ditch the aspirational "Ideal Client Profile," which represents a rare, perfect-world scenario. Instead, build a "Target Client Profile" that defines which customers will perceive the most meaningful value from your offering. This provides a realistic, operational benchmark for qualifying leads.

To create resonant content, move beyond guessing customer problems. Analyze transcripts of past sales calls with an AI tool to identify recurring pain points, common questions, and the exact language your audience uses to describe their challenges.

Instead of asking AI to generate generic blog posts, use it for strategic ideation. Prompt ChatGPT with a detailed description of your ideal client and their transformation, then ask it to list their top 25 problems or questions. This provides a roadmap for creating highly relevant, problem-solving content.

Instead of theorizing about their Ideal Customer Profile, Assembled's first GTM hire created a list of all existing customers. By looking for patterns and creating groups based on traits like tech stack (Zendesk), agent count (20-200), and channel complexity, a data-driven, highly specific ICP emerged organically.