B2B lead capture with an agent that filters

The challenge

The sales team spent its mornings reading forms and emails that mostly did not fit the service. The good opportunities were answered late.

How it was done before

Every form landed in a shared inbox. The salesperson read them in order of arrival, looked the company up on Google and judged by eye whether it deserved a reply; the good ones waited behind the bad ones.

What we built

• Every incoming contact is enriched with public information about the company.
• An AI agent scores it against the ideal-client criteria defined with the team.
• It drafts a first, contextual reply that a person reviews before sending.
• Qualified contacts land in the CRM with a summary, a priority and a next step.

Human oversight

No reply goes out without human review. The agent scores and drafts; the salesperson decides. The scoring criteria are reviewed every month against the logged hits and misses.

What the client takes away

• Agent connected to the business criteria
• CRM and calendar integration
• Lead quality dashboard

Outcome (expected)

The specific metrics — hours recovered, errors avoided, response times — are agreed at the start of each project and measured afterwards. This scenario is illustrative: we do not publish figures we cannot back up.

Limits

• The agent does not close deals or promise terms: it prepares the conversation.
• Its judgment is only as good as the ideal-client definition it is given.
• Ambiguous contacts go to human review, not to the bin.