The nine agents
Aphily structures an AI outbound team as nine agents: eight specialists that each own one stage of the outbound function, and a Partner Agent that orchestrates them and is the buyer's single point of contact.
Partner Agent
The single agent a buyer talks to when working with Aphily, sitting above the specialist agents and orchestrating them on the buyer's behalf. Instead of managing eight separate tools or dashboards, the buyer has one relationship — the Partner Agent routes requests to the right specialist and reports back in one place.
Research Agent
Turns raw materials — a website, a deck, existing customer data — into GTM intelligence within 48 hours, giving the rest of the team a factual base to target and write from instead of generic assumptions about the buyer's market.
ICP Agent
Defines ideal customers through a 34-column targeting matrix, so the criteria the rest of the team prospects and writes against stays specific instead of a loose description of "who we sell to."
Prospecting Agent
Finds and verifies prospects against the ICP Agent's definition across 200+ data sources, keeping targeting criteria and list-building tied to the same definition instead of drifting apart over time.
Copywriting Agent
Writes unique emails using 149 personalization variables per prospect, so outreach reads as researched rather than templated.
Infrastructure Agent
Manages domains, warmup, and deliverability at scale, so personalized copy actually reaches the inbox instead of landing in spam because sending infrastructure wasn't maintained.
Campaign Agent
Executes multi-channel outreach across email, LinkedIn, and calls, coordinating send timing and channel sequencing across the whole list.
Reporting Agent
Delivers real-time metrics and weekly performance reports, so results are visible without the buyer having to assemble their own dashboard from raw send data.
Intelligence Agent
Learns from every reply, feeding what worked back into how the other eight agents operate the next time — targeting, copy, and send patterns all get sharper from real replies instead of staying static.