Prospecting system.
Account-based prospecting that identifies the buying committee for each target account and feeds the downstream BDR system with enriched, contextualized leads.
the problem
Most outbound systems start with a contact list and blast it. This client needed the opposite: start with a target account, figure out who actually makes the buying decision, and research the account in depth before anyone reaches out. The BDR system needed quality upstream, not just volume.
the approach
Built on Relevance AI as the agent workspace. Took an account list as input, ran decision-maker discovery (the relevant buying committee: decision-maker, influencer, gatekeeper), then did per-account research to build context the BDR system could use for personalization downstream. Structured output as enriched account records rather than flat lead lists, so the outreach layer had more than a name and title to work with.
the outcome
Became phase one of the full outbound stack, feeding the BDR system with qualified, contextualized prospects. Gave the client's team account-level intelligence they did not have before, which shifted outreach from 'who do we have emails for' to 'who should we actually talk to.'
the lessons
Prospecting and outreach are two separate problems. Bundling them into one system is how both end up mediocre. Splitting this into a phase-one prospecting engine and a phase-two BDR engine made each layer sharper and the handoff debuggable.
stack
- Relevance AI ·
- Apollo ·
- enrichment APIs ·
- Anthropic Claude