Build the profile from customers you understand
Look at customers you serve well: their industries, the problems they solve and why they chose you. Translate common characteristics into checkable criteria such as location, company type, size and use case. Avoid treating unverified budget or demand as known facts.
Use inputs with traceable sources
Company name, website, industry, region, size and public business descriptions can inform a score. Keep the source for each field. Label unknown information as missing rather than asking AI to fill the gaps with assumptions.
Include exclusions as well as positive signals
Positive criteria identify relevant companies; exclusions remove clearly unsuitable ones. If you sell through distributors, a consumer-only retailer may be an exclusion. A high total score must not hide a failed essential condition.
Read the evidence before trusting the number
Return to the original company website and check whether evidence is traceable and recent. Separate the source statement from the model interpretation. A high score with weak evidence belongs in manual review.
Revisit the profile using actual sales outcomes
Review replies, useful conversations and opportunity progress with the team. Identify which criteria actually helped. Recalibrate when your product, market or customer mix changes.
Product reference: WinLead AI PRD V2.0. Roadmap goals are distinct from accepted capabilities. Real services require appropriate configuration.
Explore the product workflow →