Tuesdays are 3 AI updates, one hot take, one action. Fridays are 1 practical real estate agent workflow.
For two years the advice was just use AI. This month the labs quietly admitted that you need a way to control all your options and handed you a dial. Three of them shipped a way to control how hard the model thinks before it answers, and once you know it is there, your results change. Here are three examples:
1. ChatGPT added a button that makes it think harder, and a slider to set how hard. On August 6 OpenAI gave every user a Think button that pushes ChatGPT into deeper reasoning on complex questions, and gave paid users a slider to control how much effort it spends per answer. It shipped the same week free users got unlimited chats, so a lot more of your clients now have this in their pocket too. The default fast answer is fine for a quick email. For a pricing question or a contract clause, the button is the difference between a generic reply and the model actually working the problem. I’m still learning when to use it and not.
2. Claude lets you set how deep it thinks, and watch it work while it does. Claude has four effort levels, low through max, that decide how much it reasons before answering, plus a toggle that shows its thinking in a panel you can open. The effort setting is useful on its own. The toggle is the part I would tell you to use, because when Claude shows its reasoning you can get clarity on important answer for your seller. Max effort for a listing-strategy memo, low for cleaning up text you already wrote.
3. Gemini put the same dial in plain sight, which tells you where all of this is going. Google rolled out Thinking Levels in the Gemini app, a straight low-or-high choice over how much the model reasons. Three companies landed on the same answer to unpredictable AI within a few months of each other: hand the user a reasoning control. That is the real story sitting under the three product notes. The agent who knows which setting to use will out-work the one who keeps the default and wonders why the answers feel thin. The age of your ability to customize how deep AI thinks is upon us.
HOT TAKE
The reasoning dial is one the most useful AI features of the year, and it’s boring on purpose. It adds no new app to learn and no subscription to buy. It rewards the one thing that already separates strong agents from average ones, knowing when something deserves the extra effort (and tokens). Most agents will never touch it, so they’ll keep taking the first fast answer and wondering why it feels thin. Learn how to use the customization of intelligence today.
1 THING TO TRY
Take one real task this week, a pricing rationale or a tough client email, and run it twice. Send the exact same prompt first with reasoning off, then with the Think button on in ChatGPT or effort set high in Claude. Read the two answers side by side. You will feel the difference faster than I can describe it, and you will know for good which of your tasks earn the deeper setting. Friday I turn this into a 10-minute test for a practical use case.
A note from Wispr Flow which I include here b/c its a tool I swear by…
PRDs by voice. Bug reports by voice. Ship faster.
Dictate acceptance criteria and reproductions inside Cursor or Warp. Wispr Flow auto-tags file names, preserves syntax, and gives you paste-ready text in seconds. 4x faster than typing.
JOIN THE WAITLIST
My only other cohort of my complet AI Marketing System for Top Real Estate Agents is coming again at the end of the year. 2 weeks to dive deep into AI and build your 2027 marketing plan. Get on the list now at maven.com/matt-spangler/ai-for-real-estate and I’ll let you know when spots are released.
-Matt
