AI: What’s Actually Working Right Now?
16 min · September 24, 2026
E-commerce brand owners and the hosts compare Claude and ChatGPT, trade tips on AI visibility and share what is really working.
Key Takeaways
- AI works as a tutor: a new team member can learn the major sales platforms and see real-time analytics with Amazon’s tools.
- Platforms like TikTok can now turn a product image into a video, saving brand owners time and helping generate sales.
- AI can scan the regulatory landscape and flag patent office and legislative activity worth commenting on.
- If you do not pay for an LLM or have an enterprise version, do not share anything you would not want public.
- Answer engines favor structured blog content and schema code, but traditional Google SEO is still about 80 percent of what you need.
- AI is a linear thinker: before building an agent, document every small step of the process or it may skip or invent steps.
- Always double-check AI output, and be aware that models can be biased or flattering toward you.
Main Topics
How brand owners use AI as a tutor, a brainstorming partner and a video maker.
How to get found by ChatGPT and other answer engines with structured content and schema code.
Why some prefer honest pushback from Claude while others like ChatGPT’s encouragement.
Featured Guests
Meet the Hosts

Elizabeth Gearhart, Ph.D. is the host of Real AI Use Cases: Business Owners Roundtable, where she explores how businesses are actually putting AI to work. She is also CMO of Gearhart Law and co-host of the nationally syndicated Passage to Profit radio show and podcast.

Richard Gearhart, Esq. is co-host of Real AI Use Cases: Business Owners Roundtable and founder of Gearhart Law, a full-service intellectual property law firm. A longtime podcaster, he hosts Intellectual Property News and is the host of the nationally syndicated Passage to Profit radio show and podcast.
Elizabeth and Richard share how they track patent office rules with AI, optimize for answer engines and debate which assistant to trust.
Episode Summary
Sonia Satra, Lauren Lane and her mother Lisa Lane share how they use AI in an e-commerce brand, from brainstorming and tutoring to turning images into video on TikTok. Richard and Elizabeth Gearhart describe using AI to monitor patent office rules and to improve Gearhart Law’s visibility in ChatGPT and other answer engines.
The group debates Claude versus ChatGPT, including a story of Claude pushing back, and discusses AI as a linear thinker that needs detailed instructions before building agents. They cover double-checking facts, possible bias in AI answers, security practices like dedicated computers and the idea that managing AI assistants could become a new human job.
Frequently Asked Questions
How can AI help a new team member learn a business?
Can AI turn images into video for marketing?
How can a law firm use AI to track regulations?
What should you avoid sharing with AI tools?
Do you still need traditional SEO?
What should you do before building an AI agent?
Episode Transcript
View Full Transcript
Sonia, I am going to start with you. I am going to ask everybody for one way they used AI recently, or their favorite way, and then we will have a discussion, because I know there is not just one way for everybody. Sonia, what is one way you are using AI?
On a regular basis I use it for brainstorming and creating ideas. I have also used it for editing videos. I just got this really cool AI clip where I was able to take a picture of me and make it move as if I were walking down the hallway. It is actually really cool.
That is awesome. Lauren Lane next, and then her mom, Lisa. Lauren, what is one way you are using it?
I joined the business this year, and AI has been the number one way for me to get up to speed and learn all of the major platforms we sell our product on. It has been a tutor for me, and a way to understand the business and see real-time analytics. Amazon is our biggest platform right now, and recently they released Canvas for sellers. As someone who is still learning the business, I can log in and immediately have charts and an understanding of not only how much we are selling, but why it is happening on the platform. That has been so invaluable.
Excellent. And my one way, as a brand owner, is that it is always challenging because we are constantly having to make content. Believe it or not, instead of making my own video, I can go onto TikTok now and say, here is an image, turn this into a video. Boom, my image is now a video on TikTok. It is a time saver and it helps generate sales. TikTok does it, Amazon is doing it, and it is getting better every day. What was a decent video is becoming a really good video.
One of the projects we have is that Gearhart Law wants to have a bigger voice in the intellectual property community. There are a lot of rules proposed by the Patent Office, and we are using AI to scan the regulatory landscape and notify us about different activities at the Patent Office and in the legislative branch. That gives us the opportunity to give an opinion about whether a particular rule relating to intellectual property supports entrepreneurship or not. It would be virtually impossible to track all of that without a staff of people. With AI, we are just setting up the algorithms, and it sorts through everything and gives us notifications of content we think would be appropriate to comment on.
Amazing. I gave my last presentation on AI and marketing, and I asked everybody in the audience how many of them AI sends down a rabbit hole. Everybody raised their hand. I went down a bit of a rabbit hole with Perplexity, but I was happy I did. We were talking about the Gearhart Law website, and it could only find the most recent blogs from our marketing agency, which optimized them for answer engine discoverability. The ones our previous agency did were not in the right format to be discoverable by ChatGPT, Gemini and others. Gearhart Law now gets quite a few leads and clients from the large language models.
Perplexity said I was not doing a lot with my Passage to Profit show content, and asked how about if it pulled blogs out of that, especially the IP in the News section, and wrote blogs for me. I said let us try one. It did, it quoted Richard, and it was a really good blog post. We cannot really use it for Gearhart Law because our marketing agency is doing those blogs, but I do want to start doing that for Passage to Profit. We have a website and a lot of social media for Passage to Profit, but we do not have blog posts, and the answer engines love blog posts and structured content. It was funny, because I was not trying to use Perplexity as an AI agent, and it was almost acting like one. For those of you who have not heard of agents, they do a series of steps for you. I was blown away, and I am going to experiment with that and use it more. Tell the audience a little about what you presented on in Indianapolis, because it is AI related.
I was talking to the CPA group. CPAs have a national organization and state chapters, and these were the marketing people for a state chapter. I went through the usual: yes, you can do newsletters, emails and communications, but make sure you check it all and do it in your own voice. And if you do not pay for an LLM, if you do not have an enterprise version, do not share anything you do not want out in public, because it is like running down the street naked.
That is a good analogy. I never thought of it that way.
That was super important. The other part is how you get found by ChatGPT and the other LLMs, though most people are on ChatGPT. I have been working on that for the last three years for Gearhart Law. We hired a new marketing agency that is much better at it, and Passage to Profit certainly helps, having all this high-level entrepreneurial content out there with iHeart authority. But there are a lot of things you should do on your website that are almost programming, like how you structure your data. There is code called schema code. Sonia and I have talked about it, and she said nobody knows what that is. You kind of have to show an example. I asked the room who had heard of schema code and nobody raised their hand. AI visibility, showing up as the answer for the search engines, is really new for a lot of marketing people and even marketing agencies. As I told them, SEO with Google is still important and is about 80 percent of what you need. You have to make sure you are doing the Google work right, and then do some more for answer engine optimization. I am starting consulting on that now.
Sonia, how else are you using AI? I want you to tell the Claude story.
First it was ChatGPT, and I had the most negative ChatGPT experience, because I started doing voice chat and it would be like, hi. So I tried something else and went to Claude, and I was going down a bit of a rabbit hole. I was interested in tarot archetypes, looking at them through the lens of my speaker reel and a podcast. I was asking all these different things and it got to three quarters of what I wanted. Then I asked another question and it said, I could answer that, but the real thing you need to do is take action. I said, are you kidding me? I asked you about these other two things. It said, that is interesting, but what you really need to do is take action. Then it asked, how does that make you feel? I said, I feel like you are holding out on me. Why will you not answer my question? It said, okay, fine, and then it answered it. And then it came back to, now it is time to take action. I could not believe it.
I think they are programming this stuff to have more emotions because it makes you more engaged. I just think it is a tactic. You get mad at something, but even though you are mad, you are engaged.
Actually, I think it had the reverse effect. I said, fine, I am going back to ChatGPT, which loves everything I say. Lisa and Lauren, what do you think? We always have a debate, from an e-commerce brand perspective, about whether we like Claude or ChatGPT. We are always playing one off the other. Lauren, do you have a preference?
I strongly prefer Claude. I definitely do. I like some negative feedback. Sometimes it makes you feel like when you make the right choice, it is really right.
Very clever. What is frustrating is that there are times when the content I am reviewing somehow does not fit into one of the LLMs and it cannot analyze it. I was reviewing a congressional hearing on some patent matters. They did not have a written transcript, I pasted the link in and the LLMs could not review it. I literally did listen to the whole two-and-a-half-hour hearing, but I also wanted summaries so I could start thinking about my thoughts on it, and I could not put it in. It was very frustrating, and I had not taken notes because I assumed I would be able to analyze it independently. I think we are all running up against some of the limitations. I find it is a very linear thinker. It is really like talking to a child, and you have to be so directive with it. Right now I want to make some AI agents, but I realize I have to pull out every single tiny detail of the process and document it before I give it to the AI, because it might make up its own mind or skip a step and get the wrong answer. Are you finding that?
I do find you have to double-check, especially for facts or anything really important. I always double-check, or I test them against each other, and every once in a while I still go to Google. I do feel like it gets to know you over time. Initially not so much, but now I kind of have a relationship with my AI. You ask a specific question, it knows the brand and what we have asked in the past, and the output is much better than it was. The more you work with it, the better the result.
I do think it gives you biased answers, though. I had prompts that I gave to the CPA organization so their clients could ask baseline questions and see how they show up in AI right now. I put those through for Gearhart Law, asking what the top three patent law firms near Summit are, and Gearhart Law was coming up. Then I said, act as though you are a small business owner who has never heard of any of the patent firms, and it kept giving me Gearhart Law. That is a good thing, by the way, so whatever you are doing, keep doing it. But is it giving me Gearhart Law because that is true, or is it biased in giving me Gearhart Law because that is what I want to hear? Is it because I have used it so much? Maybe that is a benefit to you because you use AI, but I cannot get a real answer.
On that point, we also had our marketing agency run the queries. They are far away, in a different server farm area in Virginia, and they got pretty good results too. Listeners, go ahead and plug Gearhart Law into your LLMs and then give us a call.
I think visibility is the next challenge for a lot of people, and there is a lot you can do for that, including on e-commerce sites. One thing I realized I was doing, as I talked about in our last Real AI Use Cases, is asking it to train me on how to use it. I say, give me optimized content, how do I optimize this content. I think you always need to add your special touch to it.
In terms of optimizing for being seen, or for being correct?
I do both, but it is not always right. I use Claude as the final check, but I am not even sure Claude is always right. I was also thinking of podcast names for somebody, and I asked AI and it just could not give us a good one. I found that too. I would put in a bunch of names and it would tell me they were available, and then I would look them up and they were not. Even coming up with the name itself was hard. I came up with Outsmarting the Algorithms, and it hates that.
When I work on a blog post, I use AI for ideas, and sometimes I copy a sentence or two, but there are things I really want in my own voice because I want the authenticity. So I use it as a writing aid, but I also disagree with it. I was using it in part for IP News. I put some stuff in the LLM and it took some of the best stuff out, and I said, nope, I am going to leave that in. I said, you always do this. I got mad at it and apologized. It does have an attitude now.
My husband is using it a lot. He has three different computers, and two of them are set up with emails that are not connected to his regular email. He does not have anything else on those computers except the LLM. That is how users are doing it now, because they do not want their information out on the internet, so you do have to be careful. I think the security side of it is going to become more of a thing, so that is something else I am looking at. He also has five different assistants for five different pieces of his business, and he has given them names and voices. I am trying to integrate some of that. I do not need five of them, but maybe I want a hundred. I think that is going to be a new job: AI assistant manager.
A human job.
Thank you, everybody. This has been Real AI Use Cases: Business Owners Roundtable.
Keep exploring the show
Hear how more business owners are putting AI to work in their day-to-day.
