Learning loop

ReadyBuyerFinder learns which signals and fit rules lead to real conversations, and suggests better weights. It never changes your weights by itself.

Where the data comes from

Replies and meetings come back from OutreachSender and ReachDak (see Sequencer contract). Each outcome is compared with the signals and fit rules behind that lead.

• Positive: interested, wants a meeting, referred someone, meeting booked.
• Negative: not interested, asked to stop, finished with no reply.
• Ignored: bounces, out of office, wrong person.

When you get a suggestion

• At least 20 outcomes for the agent's ICP, with both positive and negative ones.
• At least 5 outcomes for a signal type or fit rule before its weight can change.
• Each weight moves by at most ±30% per round, and only when the change is 2 points or more.
• Up to 6 changes per suggestion, and at most one new suggestion a week per ICP.

What you see

The agent page shows the suggestion with a reason for every change, for example:

Funding: 18 → 23
Leads with a funding rounds signal: 9 of 20 replied positively (45%) vs 30% overall.

Click Accept to apply the new weights to the ICP, or Dismiss to keep yours. Accepted changes are recorded in the audit log. New and refreshed leads use the new weights.

See How scoring works for the weights themselves.