Summary4 min read

Podcast Summary: The Law Firm Marketing Minute

Episode Title: How To Influence What ChatGPT Says
Date: February 11, 2026
Host: Spotlight Marketing + Branding

Overview

This episode dives into a crucial and timely subject for law firm owners: understanding how ChatGPT and similar AI models source their information and how you can actively influence what these platforms say about your firm. The host provides both a technical explanation and actionable marketing guidance, ideal for legal professionals aiming to increase their online visibility and shape the narrative presented by AI-driven content engines.

Key Discussion Points & Insights

1. Where ChatGPT Gets Its Information

ChatGPT draws from a massive array of public data sources, including:
- Major websites: “Big public websites, government websites, CNN.com, Fox News… Wikipedia.”
  (A, 00:09)
- Forums: “Reddit, Quora, Yelp… especially Reddit.”
  (A, 00:19)
- Blog and Video Content: AI platforms can crawl videos for transcripts and include them in responses.
  (A, 00:28)
- Offline Data: “Books, textbooks, magazines, academic papers, this stuff is also loaded into ChatGPT.”
  (A, 00:41)

2. Video Content Matters

The host reassures those who prefer video over writing:

“If you don’t like to write and you really like to create video content, that’s okay. You can create content that’s still going to be found by ChatGPT.”
(A, 00:29)

3. Your Website’s Role in the AI Dataset

Your law firm’s website may end up feeding ChatGPT, but only under certain circumstances:
- “Unless you have done some things wrong, your website content is going into the Chat GPT data set.” (A, 00:14)
Implication: If you want to influence the AI, ensure your public content is accurate, robust, and present on your site.

4. Offline Content’s Influence

Not just web content matters; books, academic papers, and other offline resources shape AI output.

5. How AI Decides What to Say — The “Stacking the Deck” Analogy

Answers aren’t chosen deterministically. The process is likened to flipping over a random card from a stacked deck:

“When ChatGPT is answering your question, it is flipping over a random card, okay? ... Our goal is to stack the deck. We want to give enough content to ChatGPT that when it randomly flips over its card to decide how to answer the question, it chooses our answer.” (A, 00:48)

6. Teaching AI the “Right” Answer

The famous Dr. Seuss analogy illustrates how repeated information influences AI responses:

“That’s because Dr. Seuss has done a really great job of stacking our large language models with that phrase. ... That was the only acceptable answer. That’s what we’re trying to do.” (A, 01:12)

7. Establishing Trust Signals

Faced with so much information, AI platforms rely on “trust signals” to choose which sources are credible:

“‘Trust signals’... are going to determine what these AI search bots use to decide what information is credible.” (A, 01:40)
Implication: Legal websites need authority cues, accurate legal citations, and up-to-date information.

Notable Quotes & Memorable Moments

On sources:

“Reddit, Quora, Yelp. Okay. Really big source of information for ChatGPT, especially Reddit.”
(A, 00:19)
On content creation:

“You can create content that’s still going to be found by ChatGPT.”
(A, 00:29)
On the stacking the deck strategy:

“We want to give enough content to ChatGPT that when it randomly flips over its card to decide how to answer the question, it chooses our answer.” (A, 00:52)
On influencing AI outcomes:

“As we influence AI search engines, we are trying to teach it what the right answer is.” (A, 01:13)

Timestamps for Key Segments

00:00-00:14: Overview of ChatGPT’s data sources
00:14-00:29: Your website’s data and importance of video content
00:29-00:41: Inclusion of offline sources
00:41-01:12: The “stacking the deck” analogy and Dr. Seuss reference
01:12-01:40: How repeated content trains AI and introduction of trust signals

Actionable Takeaways for Law Firms

Publish authoritative, frequent content—written or video—on your site and on reputable public forums.
Ensure your website is technically optimized to be publicly visible (avoid unnecessary blocks or “noindex” tags).
Engage with trusted platforms (e.g., legal publications, academic sources) to increase your authority as referenced by AI.
Be consistent: The more your message appears (accurately), the more likely ChatGPT will repeat it.

This episode offers law firms a practical roadmap to shaping AI output about their services and expertise—by stacking the content “deck” in their favor, and sending clear trust signals with their public digital and offline footprint.

Loading summary

Transcript2 lines

[00:00]
A
Big question, how does ChatGPT decide how to answer question, how to how to answer questions? Like I said, it is trained for many different data sources. Okay. Big public websites, government websites, CNN.com, fox News. Right. Big media websites. Wikipedia. Right. Big public prominent websites. But our personal websites, if you have a website for your business, your. Unless you have done some things wrong, your website content is going into the Chat GPT data set. Okay? Forums are a big one. Reddit, Quora, Yelp. Okay. Really big source of information for Chat GPT, especially Reddit. Blog and video content is indexed. Okay. ChatGPT, the other platforms have the ability to crawl a video, pull a transcript out of it, and include that information and in its answer. So if you don't like to write and you really like to create video content, that's okay. You can create content that's still going to be found by ChatGPT. Right. This is interesting. A lot of people don't realize this offline data as well. Okay? Books, textbooks, magazines, academic papers, this stuff is also loaded into ChatGPT. Okay? All right, so there's all this information out there. Here's the way to think about it. When ChatGPT is answering your question, it is flipping over a random card, okay? And that card is going to have the answer. And your goal? Our goal is to stack the deck. Okay? We want to give enough content to ChatGPT that when it randomly flips over its card to decide how to answer the question, it chooses our answer. Okay, the Dr. Seuss quote, right? Red fish, blue fish, whatever. I asked you guys to spit out that answer. Everybody answered bluefish. That's because Dr. Seuss has done a really great job of stacking our large language models with that phrase. We have all been programmed that that was the only acceptable answer. Okay? That's what we're trying to do. As we influence AI search engines, we are trying to teach it what the right answer is. So how does it work? Because there's so much content out there, something called trust signals that are going to determine what these AI search bots use to decide what information is credible.
[02:22]
B
The law firm Marketing Minute is brought to you by Spotlight Marketing and Branding, where we help solo and small law firms get more clients and better clients. If you want more details, visit growmylawfirmfast.com.