Every Tab in My Terminal Is a Marketer: One Person, the Output of Ten | Jason Alafgani

One marketer matches a ten-person team using AI tabs

Jason Alafgani runs Caddi's marketing alone and says he matches a ten-person team by treating each tab in Claude Code as a specialist. The work is repeatable and context-heavy, and a stalled approach restarts in a fresh tab. Caddi sells the same idea to law and wealth firms, agents that reason with AI but act in deterministic code. Since AI makes outreach cheap, his demand strategy leans on channels others will not run and on clarity answer engines can extract.

Takeaways

• Ten AI tabs deliver a ten-person team's output. He treats each Claude Code tab as a person, one for web marketing, one for SEO, one for ads, one for direct mail, one for outbound SDR. His solo output matches the nine-person team at a previous Series C startup.
• Caddi agents think with AI, act with code. Because law firms and wealth managers need accuracy, Caddi builds hybrid agents whose actions are deterministic, so run 4,000 matches run one. Low-volume complex work can let Claude reason end to end. Caddi, two years old, sells to the Am Law 100 and top RIAs.
• Events, outbound and SEO drive demand. Caddi's best channels are events and outbound, with SEO and web driving inbound. When AI makes outreach easy for everyone, differentiation comes from doing what others cannot or will not do, and events are capped by conference capacity.
• Competition now comes from incumbents. Caddi once saw no competition because its approach was novel, and startups rarely collide. Now AI lets everyone build everything, and adjacent players are pushed by investors and boards to expand, so the threat comes from incumbents rather than startups.
• Legal AI demand now hinges on ROI. Demand in AI shifts every three months. A year before the conversation, law firms were unsure about AI and considered a single bet on Harvey. Now they have done a lot of AI work and ask where the return is.
• Answer engines reward upfront clarity. Where headlines and blurbs once hooked human readers, Alafgani says simple clarity upfront is what answer engines pull from. He calls AEO an evolution of SEO and says the data is opaque, since some tools just re-run a query through an API.
• Keep the source of truth machine-readable. Alafgani tells new marketing heads to be comfortable being uncomfortable and to keep the source of truth in a repo Claude can read instead of Google Docs or Notion. He wants predictability across SEO, outbound and events, with a fourth channel in reserve.

How a Caddi agent goes live

1. New clients build an agent and test it against live data to see what it would have done before production.