How Note Investors Are Actually Using AI (ChatGPT, Claude & More)

Published September 2026  •  Based on a Paperstac Podcast Episode  •  12 min read  •  Paperstac Education

More note investors are using AI — ChatGPT, Claude, and similar tools — in almost every part of their day now. Writing a listing, reviewing a collateral file, catching up on email, even predicting what happens next on a non-performing loan. In this episode of the Paperstac Podcast, Rick Allen and TJ Osterman walk through exactly how they use it. It’s tool by tool and task by task, including a new AI-powered feature coming to Paperstac’s platform.

Why AI Matters for Note Investors Now

AI has moved well past being a novelty. Rick and TJ point out that even borrowers are using tools like ChatGPT to draft foreclosure defense responses now. Sometimes they make mistakes by leaving in obvious AI-generated boilerplate. That alone is a sign of how mainstream these tools have become on every side of a transaction.

Rick shared a simple example: his father, a retired attorney, needed to read through a 103-page legal brief for a family member. Rick uploaded it and asked for a summary and the key issues. The tool returned a clear breakdown in under 30 seconds — a task that would have taken hours by hand.

That shift shows up across real estate broadly, not just among investors. NAR’s 2025 Technology Survey found that 41% of Realtors already use AI content tools like ChatGPT, Copilot, or Gemini in their day-to-day work.

How to Get Started, Without Overthinking It

Both hosts were clear: getting started is simpler than people expect. Pick one tool — ChatGPT, Claude, or Gemini — and just ask it a question. The free tier is plenty to begin learning.

The real shift happens once you realize AI isn’t just a search engine. It functions more like an employee you can direct. Tell it what role you want — a financial analyst, say, or an expert in land deals. Then give it context about the specific task.

Using AI for Research and Faster Learning

Research came up as the most immediate use case. Instead of running dozens of separate Google searches and combing through each result, AI can synthesize a topic into a single, usable summary.

This is especially useful for anyone new to note investing. The industry has its own vocabulary, and it’s easy to feel lost in a conversation full of unfamiliar acronyms and structured-finance jargon. That’s particularly true once you move from single notes into larger portfolios. TJ described sitting in early meetings taking notes just to look up terms afterward. Today, that same research happens in the moment — paste in an unfamiliar term or document and ask for a plain-language explanation.

That same approach works for cold outreach, too. A new investor can take a confusing email from an experienced note seller and ask AI to translate it. That makes it possible to respond with confidence. No more hesitating out of fear of “looking stupid” — TJ specifically called that fear out as one of the biggest reasons people stall out early.

Writing Stronger Note Listings

Seller descriptions were the clearest day-to-day win. TJ described being a self-described perfectionist who used to spend 30 minutes wordsmithing a single listing description. With AI, that same task takes closer to five minutes — describe the situation, ask for a clear summary, then refine it down until it’s concise.

A strong seller description matters more than sellers often realize. A listing with solid numbers but no story tends to get skipped over, because buyers want context before they ever reach out. For the fuller picture on what a complete listing needs, see our guide on writing a mortgage note listing that attracts buyers.

AI-Powered Collateral File Review

The most significant update in this episode: Paperstac’s upcoming platform release (expected by the end of the month) includes AI-powered collateral file analysis.

A seller can drag and drop an entire folder of loan documents — for a single note or a batch of ten. The system automatically identifies and labels each file: assignment of mortgage, allonge, servicing notes, payment history, insurance documents, title policy, and more.

On the buyer side, that unlocks direct questions. A buyer can ask something like whether there’s ever been a break in the chain of title on a specific property. The tool responds with a sourced answer, pointing to the exact page and document it pulled from.

Predicting Outcomes on Non-Performing Notes

During testing, Rick asked the new tool to forecast the likely outcome for a specific non-performing loan. Would it head toward reinstatement, a deed in lieu, or foreclosure? The system read through the servicing notes, payment history, and legal file, then returned a reasoned answer. It flagged, for example, that the most recent BPO showed a vacant property, which pointed toward foreclosure as the likely path.

That kind of forecasting would traditionally take an entire team of analysts. It isn’t a guaranteed outcome, but it gives an evidence-based starting point for a workout decision. See our exit strategies for a defaulted note guide for the full range of paths a non-performing note can take.

Building Your Own AI “Chief of Staff”

Beyond one-off tasks, AI can run as an ongoing daily assistant. TJ described a setup used by a colleague: each morning, AI reviews the inbox, flags the most important emails, and filters out newsletters. It delivers a spoken daily digest, all while the colleague is out on a walk.

The system can be connected to tools like email, calendars, Slack, and spreadsheets. That makes it function less like software and more like an assistant. It never gets tired, never misses a step, and can be duplicated the moment you need a second one.

AI for Lead Generation and Driving for Dollars

For real estate and land investors, AI can streamline the classic “driving for dollars” workflow: spotting vacant properties and following up with outreach. TJ mentioned once building an app for exactly this purpose years earlier. He joked that with today’s AI tools, the same project would cost a few hundred dollars to build instead of tens of thousands.

A simple version: connect a vacant-property data feed to an AI workflow. Have it draft, or queue up, mailers automatically as new vacant properties are identified. A manual afternoon task turns into something that just runs in the background.

A Few Ground Rules for Using AI Well

A few practices came up repeatedly as ways to get better results:

• Ask it to push back. By default, AI tools can agree too readily. Explicitly asking a tool to challenge your thinking — or to ask clarifying questions before offering an opinion — produces sharper, more useful answers.
• Use voice mode for brainstorming. Talking through an idea out loud, in the car or on a walk, can surface a different kind of thinking than typing.
• Give it real context. The more background a tool has about your business or a specific deal, the more useful its answers become.
• Always verify. Every use case in this episode — research, forecasting, collateral review — comes back to the same principle: check the work before you rely on it.

Key Terms to Know

Frequently Asked Questions