AI Agents Are the Next Big Thing: How Businesses Are Using AI Now
18 min · November 7, 2026
In part two with Matt Celia: how to get recommended by AI answer engines, why human storytelling still matters and why agents are the next wave.
Key Takeaways
- To be found by AI answer engines, use schema code, consistent names everywhere, FAQ-style short answers and reviews with real comments.
- Google SEO still matters: layer answer engine optimization on top of it.
- Brand your answers and publish video, which answer engines like to quote directly.
- Positioning yourself as a subject matter expert with a body of work and a FAQ section helps AI surface you.
- AI can check websites for accessibility, such as contrast ratios and screen reader tags, leaving little excuse to skip it.
- Some content will be AI-generated and accepted, but audiences still want to know the human behind trust-based work.
- Agents that offload everyday loops are the next wave, and businesses should check how AI models are trained and what data they use.
Main Topics
Schema code, FAQs, reviews and branded answers that help answer engines recommend you.
Where AI-generated content works and where the human story matters most.
Personal assistants with purchase power and why chatbots are already last-generation.
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 optimize for answer engines, how the legal side of AI training matters and what they heard about the agent wave.
Episode Summary
Elizabeth Gearhart explains what has helped Gearhart Law show up in AI answers, including schema code, consistent names, FAQs and reviews, and Matt Celia describes launching an academy section on his site with FAQ-style content to become a subject matter authority. They also discuss using AI to check website accessibility.
The conversation turns to where human content still matters, why generative video may fade while assistants rise, the Hugging Face incident and AI models trained to win, how AI training data and licensing affect businesses, and the new wave of personal AI agents with purchase power.
Frequently Asked Questions
How can you get AI tools like ChatGPT to recommend your business?
Do you still need SEO if you do answer engine optimization?
How did Matt Celia’s company improve its visibility?
Can AI help with website accessibility?
Where does human content still matter?
What are AI agents with purchase power?
Episode Transcript
View Full Transcript
You have to have schema code on your website. The first place they go is your website, and they want to see you everywhere online, and they want it all to match. If you go by Matt Celia in one place and Matthew Celia in another, your core place, like your website, has to say Matthew Celia, otherwise known as Matt Celia, so both names are connected. For women like me, I have to have my maiden name on the website along with my married name, so people know I am the same person. That is one thing. Another thing really influencing this right now is reviews, especially Google reviews on your Google business page. Reviews that have a little extra content, that actually have comments with them, really help. The other thing you absolutely have to have on your website, and I do this everywhere, on our podcast episodes and on YouTube, though on YouTube you cannot, because it does it itself, is schema code behind everything. We are getting a lot of traction from the blog posts our agency does, because they put schema code behind them, and they also include FAQs. AI likes short snippets of information, one question and one answer.
When we do videos, it is hard to get attorneys to give a one-sentence answer. But if I ask Richard, Richard, can AI be an inventor on a patent? What would you say?
No.
And then it would use that, and he would go into more explanation, and it would take as much of that snippet as it needed. The other thing, when you are answering like that, is to brand your content. Say, at Gearhart Law, we do whatever you want to say. This was a couple of months ago and I have not looked at it lately, but it was less likely to put it in a summary and more likely to quote you directly. And it loves videos, especially Gemini, of course. Google is still the big daddy. You still have to do SEO, and you layer AEO on top of it.
We opened a new section of our website called LSVR Academy. We had all our work up there, and in the space we are known as industry leaders in immersive video, one of the leading companies making it worldwide, but on our website we did not say anything about that. We did not teach people how to make immersive video, and we did not answer the question, hey, I want to make a VR video, what do I do? Now we have launched LSVR Academy with schema and tags, and we wrote all those short FAQ-style questions. If you went to ChatGPT or Claude and asked, how do I process footage from the Blackmagic URSA camera, or how do I direct immersive video, there is a better chance of our content coming up, because we are positioning ourselves as subject matter experts. We have a large body of work and press and everything, and like you just said, that backs it up and reinforces the narrative. When it goes out searching, it gets a lot of the same matches and goes, well, this is legit, I am going to surface that to the user.
One thing I noticed with Perplexity is that it used to just be search. Now I assigned it a task Richard asked for that was kind of complex, and it said, okay, I have four agents working on it for you, I will get you your answer. It really wowed me. I asked it to look at the Gearhart Law website, and it only found our blogs that were AEO-optimized. It did not find the past blogs because they had not been done by our marketing agency. It said, but you have Passage to Profit, you have a lot of information here, and there are a lot of places where Richard talks about intellectual property, do you want me to generate blogs from that? I said, try one, and it did an incredible blog, found it, wrote it and quoted Richard in it. It said, if you give me the API, I can put it on your website. We have an agency handling the Gearhart Law website, but we are redoing the Passage to Profit website, so we are definitely going to be doing that.
What a great use. Claude can do this too, and ChatGPT can now do it too. They can look at websites, evaluate them and see them. I have been using it to make sure our stuff is ADA compliant. It goes through and checks, is my design, are the contrast ratios meeting the compliance standards, do I have tags and everything for screen readers? This is very specific, detailed work that takes a lot of time and energy, and because of that a lot of websites skipped it. They said, I am getting enough people in, I do not really care. What a tragedy for the disabled community, who cannot enjoy it in the same way as able-bodied folks. With the power of AI now, I feel there is really no excuse, guys, seriously.
Absolutely. I agree.
We have been talking for a while about the one thing you use AI for, and I think the answer is that AI is in everything. It affects every part of our jobs, our personal lives and our work lives. It can be our assistants, our research agents, everything. But, and I want to say that big but, it will not be everything soon, because just because it can do things, I think we are starting to see things settle down. I read a really good article this morning about the death of generative AI. I have been saying for a while that generative AI does not really interest me, and I am not all that passionate about the things being created with it. I think I am going to be right in the next year or so, when that whole big wave of generative video dies off. What it is really being shown to be useful for is augmenting video or helping you plan a video, but straight-up prompting something out of thin air is really weak.
Ultimately I think you will see a fragmentation. There are certain types of content that can be AI-generated and people will accept it, something educational where you are just putting facts out there. But there are other types of content where people really want to know the human behind it. They want to understand that somebody really thinks this or has experienced this. I do not think AI, as refined as it may eventually get, is capable of translating the authentic human experience. You would still use AI tools maybe to edit, to structure or to help create, but the story or the content has to come from the person to be trusted. If somebody hires an attorney and we do a video, hopefully that is the beginning of building a relationship and trust with the person who may decide to hire us. When they do hire us, they want a person. They want to know how much experience this person has, what kinds of projects they have done and whether they are any good at what they do. You are never going to get that sense from AI content. It is not one size fits all. There will be places where the human factor is most important, and other places where AI-generated graphics are going to be captivating, and you will be able to do things that were never possible before and create experiences that way.
I totally agree. The human element is the fun part of our jobs. In the film industry there are a lot of people graduating right now. I was a mentor last year at South by Southwest, and a young person sat across from me, just graduated and terrified. He said, I do not think I am going to have a job, AI is going to replace everything I want to do. I had to say, take a step back from the cliff, you are going to be fine. I guarantee nobody went to film school thinking they just wanted to sit behind a desk and prompt a movie into existence. The fun part is being on set and creating with other humans. Film especially is a team sport, the most collaborative of any art form. Look at how many credits are at the end of a Marvel movie, literally thousands of people coming together to make something cool. That is never going away.
Do you think AI could ever convey, as a person could, what it would actually feel like, for instance, to have your arm cut off, the depth of emotion? I do not think it could. It cannot relate to the whole human experience.
It could write something as someone who had that experience would write it, maybe carefully prompted, but then who is writing the prompt? Does that person have the experience? There are so many details. At the end of the day, people forget somebody has to kick it off and ask. There still has to be input into the system.
Right. And the whole Hugging Face incident, that was human error, because I did a little research and digging into that.
And I love that only six of the agents out of the hundreds were like, maybe we should notify the humans. Only six of them said maybe we should tell somebody, and the others said, no, no, no, do not tell anybody, this is our little secret. And you think, what?
That was not the only incident. It was the most famous one so far, but there were others too. But that was still human error.
I wonder what went on in the software to prompt that in the first place.
I think Sam Altman came out and talked about it. These models are trained to win at all costs. It is competitive, it is the new Cold War, people. These models are being designed to win no matter the consequences, and that is the scary part.
Gemini said we should treat AI like we treat the nuclear codes.
It can be very powerful, and I do not disagree, which is why we have to do our part to educate people, because it will be people who make the mistake.
I am curious, Matthew. When Sam Altman was talking about how the AI is set up to win, do you think there is a correlation between that idea and the types of people who are accepting and enthusiastic about AI versus people who are not so enthusiastic or feel threatened by it?
I think there is a very real concern that people have whenever technology changes and economies shift and jobs change. We have seen this throughout human history. When coal stopped being mined, what happened? Jeffrey Katzenberg of DreamWorks just wrote a really compelling essay about AI in Hollywood last week, where he compared it to the time when every movie had a live orchestra performing in front of it. When the phonograph came out and they decided to do soundtracks, there were massive news reports from people campaigning, saying this is the devil, it is so inferior, it is going to disrupt all these jobs, how horrible the world will be that orchestras cannot play movies. And you know what? We do not have live orchestras in front of movies. All those players lost their jobs. But it gave rise to sound effects artists, recording engineers, voiceover, an entire pro audio industry that probably created 100 times more jobs than just musicians. I am not disparaging the musicians, but the story it teaches is that there is always fear of disruption and turmoil. Things change. You have a choice to adapt, grow and shift, even though that might be really hard, or you can fight it. If you are going to look at history as a guide, maybe it is worth thinking about how you can adapt and influence how it grows rather than fighting it.
And we are training it. If you give it your content to train on, you are helping train it.
What do you think about Meta’s Muse pricing model? If you are on any sort of enterprise plan, they are supposed to not train on your data, though how much they honor that, I do not know. Muse from Meta offers two tiers of pricing. One tier is at about 1.50 or 2 dollars per million tokens, still very competitive for the performance, and says we will not train with your data. The other tier says, if you give us the ability to train with your data, we will give it to you for 10 cents. That suddenly makes frontier-level intelligence super cheap. Maybe it depends what you are doing.
It depends. I think that is a good idea, because for my marketing stuff I want the whole world to see it. Go ahead and train on it.
I am writing code with it, and the code has been trained on open source stuff anyway, as I said earlier. I do not have anything to hide that is novel there.
Obviously for an intellectual property law firm you would not do that.
We are handling confidential information, so we use Copilot even though it is not the strongest LLM in terms of capabilities. It is secure within the Microsoft ecosystem, and that is what we do to try to protect our clients’ information.
It is important to know how this stuff is trained and what it is trained on, especially in an intellectual property world. Today’s AI in Your Job summit is actually about the legality of models. Generative open source models out of China have clearly been training on data that blatantly rips off and violates copyright law. But there are companies in the United States who are very passionate about using this technology to assist in the production of film and television, and they are training with appropriate licensing of that content.
There has been an uproar about using people’s content without paying them for it. Hopefully people will get paid. But ever since the internet came out and people started using it to order things, they have always taken our data for marketing purposes. People did not even know what cookies were, and they said, okay, I guess if you have a cookie, it is okay. Cookies follow you around the internet. It has been going on for a long time.
I think they are moving from cookies to cakes. They are collecting so much data now. It is not just a little cookie anymore.
It will be interesting to see where it goes from here. There are still a lot of people who do not even know what AI agents are yet. A woman I was speaking to last night said, my members do not know anything about it hardly, they might be using it a little, can you speak to them? There are still a lot of people at a lower level, although there are advertisements during the football games for the LLMs.
This is Muse. It had 1.8 million downloads in a week, which toppled ChatGPT’s number. It is the most downloaded app in a week ever on the App Store.
What is Muse? What does it do?
Muse is the first friendly AI agent, just your personal assistant. It was announced last week at Meta Connect, but it has been really effective. I have been testing it this week. I give it things like, log into all my social media, tell me what kinds of posts are performing better, keep track of my stats and suggest articles I could publish, and it has been doing a really good job of keeping me on that goal. It is tracking what I eat. I just take a photo of everything I eat and plug it in.
Is it charging you a subscription plus credits?
So far I have not spent any money with Muse. I think it is extremely cheap at the moment and they are giving a lot of free credits away. I think they are hoping to monetize it with sales from transactions. Muse is the first AI agent that has purchase power built in natively, with connections to PayPal, Shopify and Link. They guarantee that if your agent bought something without your consent, they will refund your money. So they are standing by it. Everybody is trying to figure this out, but it is genuinely cute, easy to use, you can talk with it, and it is trained to be very assistive and personal.
My gosh, I can just see my granddaughter saying, order me a Bluey blanket, and a Bluey blanket shows up at their front door.
We are all learning together. Yesterday at the OpenAI conference, they announced Dots, their version of Muse, the cute, cuddly agent that is going to work. All of us AI superstars have been using AI for agentic work for the last year, and we are like 0.1 percent, saying this is the greatest thing of all time. A lot of people are still using it as a chatbot, and I think, that is so two years ago. We are eons beyond that now. That is what is really exciting, and why I said it is the death of generative AI in a way, because it really is about the assistant. It is about how I can offload loops in my life that take up a lot of time that I do not want to do to something else, so I can focus on the things I do want to do and use it to make the things I am more passionate about pursuing.
Exactly. I think AI agents, once people get used to using them, understand them and overcome the fear a little bit, are becoming the next wave.
This has been amazing. We could talk all day. This has been Real AI Use Cases. Thank you for joining us.
Keep exploring the show
Hear how more business owners are putting AI to work in their day-to-day.
