The Paperless Agent

Three coaches, three very different superpowers. Here is my take.

Brian Buffini — I worked with Brian from 1999 to 2010, so I can speak to this firsthand. His superpower is his ability to read where the market and the industry are heading and translate that into clear, actionable guidance. His Bold Predictions series is some of the best industry commentary available. High integrity, high character, and a genuinely useful read on what is coming.

Tom Ferry — I do not know Tom personally, but his superpower is clear from watching his work: network. His ability to pull together world-class experts across marketing, AI, prospecting, team leadership, and brokerage (and deliver that expertise to his audience) is what sets him apart. The caliber of people he brings together is consistently excellent.

Tim and Julie Harris — Their superpower is finding the opportunity in any given market and backing it up with practical scripts, dialogues, and tactics for actually going after it. Practitioner-based, specific, and consistently motivating.

Watch the full video: https://youtu.be/4zWv

Custom GPTs Are Going Away on December 11, 2026: Here Is What You Need to Know

OpenAI just announced the end of custom GPTs. December 11, 2026 is the date. After that, they are gone across all ChatGPT plans.

If you are using custom GPTs in your real estate workflow, here is what this means for you:

If you built your own custom GPT: ChatGPT has a migration path. You will need to move your custom GPT over to the new plugins format. The instructions are linked below and you’ll need to start this before October 26 when creation of new custom GPTs is planned to end.

If you are using someone else's custom GPT from the marketplace: This is where it gets complicated. You are dependent on whoever built that GPT to migrate it to a plugin. You may not have a way to contact them. If they do not migrate it, it disappears on December 11.

A heads up to note: custom actions inside your GPT do not carry over automatically, those need to be rebuilt by hand. And public sharing settings do not transfer either.

Get ahead of this now. List every custom GPT you rely on, check whether it is yours or someone else's, and start the migration process before the October 26 deadline for new GPT creation.

Full migration guide from OpenAI here: https://help.openai.com/en/articles/20001519-custom-gpt-retirement-and-migration-faq

🔗 https://thepaperlessagent.com
🔗 AI Studio: https://thepaperlessagent.com/ai-studio

Chris A. Scott is a Real Estate Digital Marketer & AI Strategist at The Paperless Agent.

This is the question I get asked most often right now. Here is my answer.

For writing, marketing, and client communications: Claude.

Claude is a better writer. Better at creating emails, marketing copy, and client-facing messages. Better at formatting polished professional documents. If you are doing anything that involves putting words in front of a client, Claude is the tool to reach for.

For analysis, routine tasks, and image generation: ChatGPT.

ChatGPT handles analysis and repetitive workflows well, and it is the clear choice when you need image generation. Claude's image capability is limited by comparison.

For real estate agents specifically, I lean heavily toward Claude. The ability to produce strong marketing, handle client communications, and format clean professional documents covers the majority of what most agents use AI for every day.

Learn more about using Claude in your real estate business at The Paperless Agent:

🔗 https://thepaperlessagent.com
🔗 AI Studio: https://thepaperlessagent.com/ai-studio

Chris A. Scott is a Real Estate Digital Marketer & AI Strategist at The Paperless Agent.

If you are switching from ChatGPT to Claude, you do not have to start from scratch.

Claude has a built-in feature that lets you import your memories from ChatGPT directly. Here is how it works:

Go to Claude → open Memory settings → scroll down to "Import memory from other AI providers" → click Start Import → copy the prompt it gives you → paste that prompt into ChatGPT → copy ChatGPT's response → paste it back into Claude → click Add Memory.

That is it. You cannot bring over every project or conversation you have had in ChatGPT, but the critical memories (how you like to work, your preferences, your context) transfer over and give you a meaningful head start in Claude.

If you are building your real estate AI workflow around Claude, this is the first step worth taking.

Learn more about building AI workflows for real estate at The Paperless Agent:

🔗 https://thepaperlessagent.com
🔗 AI Studio: https://thepaperlessagent.com/ai-studio

Chris A. Scott is a Real Estate Digital Marketer & AI Strategist at The Paperless Agent.

One of the most useful things Claude did during this build was ask clarifying questions before generating anything.

A lot of the time when we sit down to prompt an AI tool, we start talking and figure out what we want as we go. That works, but it often means the first several outputs miss the mark while you iterate toward what you actually wanted.

In this case, Claude recognized that I was trying to build personal software and thought ahead about what it needed to know before starting. Who goes into the CRM? How will interactions get logged? What does the follow-up cadence look like?

Those are not technical questions. Anyone could answer them. But having to answer them before the build started meant the specifications were clear from the beginning, which saved a significant amount of back and forth later.

Watch the full video to see the complete build and what I would do differently.

Full post: https://thepaperlessagent.com/blog/i-used-claude-to-build-a-personalized-crm-then-i-tried-using-it

Chris A. Scott is a Real Estate Digital Marketer & AI Strategist at The Paperless Agent.

The first prompt was straightforward: "I want to create a simple CRM tool."

Before getting into any of the specifics, I asked Claude to propose a database architecture first. That step was worth doing because it gave me a clear picture of how complicated the build was actually going to get before any data went in.

From there I gave it the core requirements: an Excel database stored locally on my computer, connected to my email and iMessages to automatically log interactions, with no manual data entry required.

The spirit of personal software is that you do not need to think like an engineer to build it. You describe what you want, answer questions as they come up, and the tool takes shape.

Watch the full video to see how the build progressed and where it hit a wall.

Full post: https://thepaperlessagent.com/blog/i-used-claude-to-build-a-personalized-crm-then-i-tried-using-it

Chris A. Scott is a Real Estate Digital Marketer & AI Strategist at The Paperless Agent.

Every live training I do, someone asks the same question: what is the best CRM?

My answer has always been the same: the one you are already using. The switching cost is almost always too high to justify starting over with a new system.

But that answer never felt satisfying, because the real problem is that most CRMs are built for the average of thousands of users and not for you specifically. So I decided to test a different idea.

What if I used AI to build a CRM that was exactly the way I would want to use it?

I spent a weekend going back and forth with Claude. The total build time was a few hours. I was pretty happy with what I came up with.

Watch the full video to see what the CRM looked like and why I ultimately stopped using it.

Full post: https://thepaperlessagent.com/blog/i-used-claude-to-build-a-personalized-crm-then-i-tried-using-it

Chris A. Scott is a Real Estate Digital Marketer & AI Strategist at The Paperless Agent.

The CRM connected to my Gmail and iMessages and automatically logged interactions with contacts in my professional network (no manual note entry required).

Based on a quarterly touch schedule I set for each person, it surfaced who I was falling behind on and recommended follow-up activity. When you drilled down on any one contact, you could see the full history of every logged interaction, sourced directly from email and messages.

The database worked. The data was clean. The interface did exactly what I wanted.

Watch the full video to see where it broke down why I stopped using it despite all of that.

Full post: https://thepaperlessagent.com/blog/i-used-claude-to-build-a-personalized-crm-then-i-tried-using-it

Chris A. Scott is a Real Estate Digital Marketer & AI Strategist at The Paperless Agent.