Alpha Apex inbound engine

Alpha Apex Group is a US consulting and recruitment firm, and one of inBeat’s sister brands in the Fieldtrip group. They got heavy inbound, and all of it landed in one place: people looking for a job, companies wanting a consultation, firms needing a search. Each type needed a different response, and a human had to read every one first.

The SDRs spent most of the day filtering spam, classifying leads, and routing them. By the time a real lead got a reply, the moment of interest had passed. The instinct was to hire more people to work the inbox. The actual problem was that no system existed to do the repetitive part.

What I built

An engine that handles every inbound lead from email and phone, end to end.

An LLM classifier sorts each lead into its real category (candidate, search opportunity, consulting, capital raise, partner outreach) with a confidence score, and filters spam before it reaches a person. A deterministic routing layer then sends each lead the right way: candidates onto their own track, search work prioritised, consulting matched to the correct internal team or external partner.

The routing is deliberately not an LLM decision. Classification is fuzzy and benefits from a model; routing is a business rule and benefits from being predictable.

From there the engine drafts a tailored first reply, posts a Slack notification carrying the lead’s context and one-click action buttons, and writes every decision back to the CRM. A conversational agent handles follow-up commands in the Slack thread, so a rep can reroute a lead by asking. Global error handling means a failure alerts the team instead of silently dropping a lead.

Result

The SDRs got back more than 6 hours a day of filtering, sorting, and routing. Revenue grew 45% over the first three months, because leads were handled at the point of interest rather than hours later.

I built it in about three weeks.