{"id":7058584,"date":"2026-09-10T15:40:47","date_gmt":"2026-09-10T15:40:47","guid":{"rendered":"https:\/\/peraltafinancing.com\/uncategorized\/stop-asking-start-assigning-john-lees-ai-blueprint-for-agents\/"},"modified":"2026-09-10T15:40:47","modified_gmt":"2026-09-10T15:40:47","slug":"stop-asking-start-assigning-john-lees-ai-blueprint-for-agents","status":"publish","type":"post","link":"https:\/\/fivemor.com\/?p=7058584","title":{"rendered":"Stop Asking, Start Assigning: John Lee’s AI Blueprint for Agents"},"content":{"rendered":"


\n<\/p>\n

\n

AI advice tends to land in one of two buckets: it\u2019s either so technical that your eyes glaze over, or it\u2019s so obvious that it becomes redundant. Fortunately, New York Times<\/em> bestselling author (and self-described non-tech guy) John Lee<\/a> knows how to split the difference. He recently sat down with Jason Abrams on the MREA Podcast<\/a> to discuss how to make AI work for your real estate business.\u00a0<\/p>\n

Spoiler alert: most people are using AI like a search bar, when they should really be using it like a new hire.<\/p>\n

Take the \u201cL\u201d Out of It<\/strong><\/h2>\n

Lee\u2019s own story started far from Silicon Valley. As a dishwasher at his mom\u2019s Chinese takeout, then an animator turned entrepreneur, his education in business wasn\u2019t exactly \u201ctextbook\u201d. But along the way, a mentor gave him a piece of wordplay he\u2019s carried ever since: Drop the L from \u201clearn\u201d and you get \u201cearn.\u201d<\/strong>\u00a0<\/p>\n

If you\u2019re not learning, you\u2019re probably not earning. Sure, it\u2019s a simple frame, but the reality is that the agents who get the most out of AI aren\u2019t the most technical. They\u2019re the most curious.<\/p>\n

Find the 4%, Not the Hours<\/strong><\/h2>\n

Before the conversation ever got to AI, Lee laid out how he thinks about effort in the first place.\u00a0<\/p>\n

While a lot of people understand that 80% of your results come from 20% of your causes (aka the 80\/20 rule), Lee pushes it a layer deeper. \u201cLook at the 20% of that 20%\u201d he said, \u201cand you\u2019ll usually find a slim slice of activity (he puts it at around 4%) driving something like 96% of your results.\u201d The takeaway isn\u2019t \u201cwork harder.\u201d It\u2019s that the agents getting the most out of their time have already found their 4% and built everything else around protecting it.<\/p>\n

Which is helpful framing when you consider that AI is really just about enabling you to spend more of your time on the 4%, and less on everything else.<\/p>\n

The Mistake Everyone Makes<\/strong><\/h2>\n

Ask most people what they do with AI, and they\u2019ll describe a glorified Google search. \u201cThat\u2019s the whole problem,\u201d Lee said. The fix is something he calls \u201creverse prompting,\u201d where instead of asking AI questions, you ask it to ask you<\/em> questions, surfacing the blind spots you didn\u2019t know you had. It\u2019s about treating AI less like an oracle and more like an employee. It\u2019s something you delegate specific, monetizable tasks to, not something you consult.<\/p>\n

He organizes it into five skills worth actually learning:\u00a0<\/p>\n

    \n
  1. How to talk to AI in specific, non-generic prompts.<\/li>\n
  2. How to spot the repetitive tasks worth automating.\u00a0<\/li>\n
  3. How to \u201cvibe code\u201d simple tools (without knowing how to program).<\/li>\n
  4. How to build a knowledge base out of everything you\u2019ve already said and written.<\/li>\n
  5. How to connect your tools together (Model Context Protocol or MCP integration) so a summary turns into an actual task landing on someone\u2019s desk.<\/li>\n<\/ol>\n

    Put It to Work<\/strong><\/h2>\n

    If you\u2019re an agent starting from zero, here\u2019s where Lee\u2019s advice gets concrete.<\/p>\n