Lead & Account Scoring

ICP & Persona Definition Agent

ai

Builds a defensible, origin-tagged Ideal Customer Profile and refreshed buyer personas by contrasting closed-won against closed-lost accounts, producing a versioned ICP with traceable attribute origins and a diff against the prior version.

Sales / Revenue OperationsCompetitive / Product MarketingSales Leadership
Versioned
ICP with traceable attribute origins
8x
Faster ICP refresh vs. manual analysis
Win/Loss
Grounded in closed-won vs. closed-lost contrast
Use Case 1

ICP & Persona Refresh

Trigger: Annual planning or a strategy shift

1.Retrieve accounts, opportunities, current ICP, and personas
2.Segment closed-won vs. closed-lost by firmographics
3.Identify attributes that separate wins from losses
4.Origin-tag each attribute by evidence source
5.Generate a versioned ICP record
6.Produce a human-readable diff against the prior version
7.Surface conflicts between data and assumptions

A defensible, versioned ICP that downstream agents consume—with every attribute traceable to its source.

Use Case 2

Assumption-vs-Data Reconciliation

Trigger: Win/loss data contradicts a held belief

1.Label the data-derived finding and the hypothesis separately
2.Present both without overriding strategy
3.Flag the tension in the diff for review
4.Avoid silently resolving the conflict
5.Hold the version pending leadership decision

Leadership sees where evidence and assumption diverge before committing the ICP.

Integrations

CRM (Salesforce, HubSpot)
ICP & Persona Records
Win/Loss & Opportunity Data
External Firmographic Data
Google Workspace (Sheets, Drive)

Industries

all

See It In Action

A sample conversation with this agent

Refresh our ICP for the new fiscal year.
Pulling accounts, opportunities, the current ICP, and personas. I segment closed-won vs. closed-lost by firmographics—segment, size, region, environment—and identify the attributes that actually separate winners from losers, not just describe customers. Each attribute is origin-tagged: 'derived from win/loss data,' 'founder hypothesis,' or 'external source.' Output is a new versioned ICP with a human-readable diff against the prior one. I'll surface anything that contradicts a current assumption.
The data says enterprise wins more, but leadership believes mid-market is our ICP—how do you present that?
I present both, clearly labeled by origin. The win/loss data shows a higher close rate in enterprise, tagged 'derived from data.' The mid-market focus is tagged 'founder hypothesis.' I don't override leadership's strategy or bury the tension—I make it explicit so the decision is informed. The diff highlights it as a conflict for review rather than silently picking a side.
Which attributes moved most since the last ICP?
Three notable shifts in the diff: EHR environment jumped from a weak signal to a strong win predictor (enterprise EHRs correlate with wins); company-size band tightened—the 500–2,000 employee range now wins disproportionately; and a prior 'high-growth' attribute weakened and is downgraded. Each shift is traceable to the underlying win/loss cohorts. Downstream prospecting, lead scoring, and audience agents will inherit these once you approve the version.