Your New COO is a Button in Gemini

The pace of AI development is relentless and as of May 2025 we've entered the era of "thinking models"—with Google's Gemini 2.5 and OpenAI's o3-pro leading the way. They’re designed to analyze, reason, and run entire business systems.

This elevates your AI from an assistant you delegate tasks to, into a strategic partner that can manage core functions. Yesterday, in the second to last class of my course, "The AI Marketing System for Real Estate Agents" I showed the students how turning on the Deep Research button inside the chat helped transform the strategy for an agents 3-month Summer campaign from good to great.

The conversation is no longer about saving minutes; it's about automating workflows. And from testing all three new models the past month, Google's Gemini 2.5 Pro, has quickly closed the gap in the race to be the COO of choice.

Why Should You Care?

Think about your typical Tuesday afternoon juggling…

• Dozens of tabs with MLS data

• Local market reports

• PDF of zoning regulations

• 3 different email threads with clients

You have all the information, but no time to connect the dots. Your brain is the bottleneck.

The "Deep Research" capability unlocked by these new models is the difference between a calculator that can add numbers and a data analyst who can tell you what the numbers mean. By processing and synthesizing massive, disconnected datasets at once, it moves beyond simple answers to deliver genuine strategy. And it adds comprehensive online research to give you an analytical edge that was previously impossible without a team of researchers.

You are the experts for your clients and this gives you a leg up in every possible conversation.

Real Estate Use Cases That Will Transform Your Day

📊Market Intelligence

• Instead of: Manually pulling comps and eyeballing a price.

• Now: Analyze this MLS export of 50 recent sales. Identify the top 5 comps for 123 Oak Street, create a pricing strategy with a data-backed justification, and chart the neighborhood's sale-to-list ratio over the last six months.

🤝Client Preparation

• Instead of: Frantically searching your email for notes before a client call.

• Now: Review my entire email history and all documents related to the 'Smith Family' buyer profile. Summarize their key priorities, list their top 3 concerns from our last conversation, and generate five tailored questions for our meeting today.

📍Neighborhood Expertise

• Instead of: Giving a general summary of a neighborhood.

• Now: I've uploaded a local news article about a new development, the school district's annual report, and a PDF of the community plan. Synthesize these to create a 'Neighborhood Briefing' for a family with young children, focusing on kid friendly options, investment potential, and community growth.

📂Deal Navigation

• Instead of: Spending an hour reading a dense inspection report to find the key issues.

• Now: Analyze this 50-page inspection report. Identify the top 3 most critical and costly issues, summarize them in plain English for my client, and draft an email to the seller's agent to open negotiations on these specific points.

These workflows, which leverage AI's ability to reason across your documents, emails, and content, were IMPOSSIBLE just 2 months ago. They represent a fundamental change in how you can run your business. If you want to learn more you can check out OpenAI’s page that gives you best practices for these new “reasoning models”.

So which one do you use? Well, first you should check with your company to see which LLM your IT team supports. For Compass, it’s Gemini, but I encourage all agents to experiment with multiple models. Pros no longer ask "Is AI useful?" but "How do I combine AI tools to excel?" I believe building fluency across a personal "stack" of tools—using Gemini for data analysis, o3-pro for creative writing, and Claude for complex problem-solving—will allow you to optimize the model for the right part of the workflow.

Reply with the first data rabbit-hole Gemini should chase for you.

—Matt