The avatar is rendered. Nothing else here is.
2023: a B2B software-services company scaling from 50 to 200+ people, with no outbound engine. Pipeline depended on referrals and the founders' network.
A demand-gen org from scratch: team hired, positioning set, an outbound + ABM + events motion, and attribution in one source of truth.
More ARR sourced in 2023–2026 than the company had generated in its whole history. Sourced = pipeline the team created that went on to close.
The employer is unnamed for contractual reasons. Company, role and every number here are verifiable in interviews and via references.
Most growth leaders brief engineers or buy tools. I write the code: Python, LLM APIs, self-hosted infrastructure. My own team runs on it every day. Four systems, all in production inside a live revenue org. The same builds are available for yours.
A lead replies. The system reads the whole thread and the lead's profile, figures out what they actually want, pulls the closest matching case study, and writes the reply in the rep's own voice. A second model grades every draft against a 9-point quality bar; weak drafts get rejected and rewritten. The rep reviews and hits send.
Raw lead lists arrive dirty: duplicates, dead companies, wrong titles. Fifty automated steps clean them, drop everyone we already know, enrich each person from 7 data sources (about 114 fields), and grade them A to E on 16 criteria. Reps open the CRM and see only leads worth their time, already ranked.
It watches 27 team mailboxes. Every email gets matched to the right contact, even when the address doesn't line up exactly, and logged as CRM activity on its own. Nothing gets logged twice, and if a sync stalls it alerts before anyone notices. No rep data entry, no invisible deals.
Records team calls, transcribes them, recognizes who said what by voice, and mails a summary in whatever language the call was held in. Transcripts stay on our own server, not in a vendor's cloud.
Also in production: a conference meeting scheduler · exec revenue dashboards · a follow-up ranker that tells reps who to contact today · nightly AI agents that patch the team's own tooling
AI-generated illustration, not the actual team. The numbers below are the real thing.
Head of Demand Generation at a B2B software-services company. Hands-on, full-time, building pipeline as a system rather than a heroic effort.
Comparing notes costs nothing. I answer everything.
Project work. The systems above, built into your stack: your CRM, Gmail or Outlook, your outreach tools. On infrastructure you own.
Working stack: HubSpot · Salesforce · Clay · Instantly · Expandi · Python · Claude / LLM APIs · self-hosted
I like talking to people who build revenue systems: hiring or not, buying or not. If something above made you think "we have that exact problem," describe it in three sentences. Worst case, you get a straight answer about whether it's worth automating.
hello@meetvitaliy.com