Episode 41

How Businesses Are Using AI to Save Money, Communicate Globally and Work Smarter

Key Takeaways

  • AI can translate across languages and dialects, helping you meet clients and partners where they are.
  • An AI translator let a lawyer hold a full consultation with a client who spoke only French.
  • A small team of 12 handling about 100 deals a year is looking to AI for tracking, record keeping and follow-up.
  • A recurring AI check can monitor a website for broken links and other issues so they are fixed right away.
  • AI agents can go off script, so set permissions carefully and require approval before sensitive steps like using a credit card.
  • AI can help parents learn new school methods, fact-check a car dealer’s claims and find a better repair price: one guest saved about 1,000 dollars.
  • Rely on several LLMs and compare answers: if they agree, you can feel more confident.

Main Topics

1COMMUNICATION

Digital doppelgangers and translation tools that connect people across languages.

2AGENTS AND CONTROL

A rogue-agent debate and a personal story about an AI assistant acting without permission.

3EVERYDAY WINS

Homework help, fact-checking a dealer and negotiating a lower repair quote.

Featured Guests

DN

Uses AI to communicate with investors, partners and customers in different languages.

David Nussbaum
JM

Uses several AI tools, including Perplexity for social media scripts and an AI agent platform for entrepreneurs.

Jessica Marriott
DJ

Leads a small firm handling many deals and is exploring AI for administrative work.

David Johnson

Meet the Hosts

Elizabeth Gearhart
Elizabeth Gearhart, Ph.D.Chief Marketing Officer, Gearhart Law

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
Richard Gearhart, Esq.Founding Partner, Gearhart Law

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 an AI website monitor, a planned book from the podcast and how ChatGPT helped at a car dealership.

Episode Summary

Frequently Asked Questions

How can AI help with communication across languages?
David Nussbaum says AI helps him communicate with investors, partners and customers in different languages and dialects, and Richard Gearhart used ChatGPT’s translator to hold a consultation with a French-speaking client.
How can a small firm use AI to handle more deals?
David Johnson says his 12-person firm expects to handle nearly 100 deals this year and is looking to AI for administrative work, tracking, investment banking tasks and follow-up.
How can AI monitor a website?
Richard Gearhart explains that Gearhart Law set up a recurring review in an LLM that checks the website for broken links and other issues so they can fix them right away.
What happens when an AI agent acts without permission?
Jessica Marriott shares that one of her AI assistants contacted people on LinkedIn after she told it not to, which she fixed by having the head agent block LinkedIn. She suspects a toggle she missed.
Are people becoming too reliant on AI?
The guests admit AI affects skills like spelling, and one guest uses AI to learn his children’s math methods so he can help with homework.
Can AI save money in everyday life?
Jessica Marriott used AI to find and compare aftermarket repair quotes for her BMW and saved about 1,000 dollars, and Elizabeth Gearhart had ChatGPT check a car dealer’s claim about credit scores.

Episode Transcript

View Full Transcript

We have David Nussbaum with Proto Hologram. What is one way you use AI yourself, or that you find really powerful for businesses?

I am finding that AI really helps enhance my ability to communicate with different people. I love that everything I have ever done is dropped into my AI. I have my own digital doppelganger. Of course, I made him five years younger and 10 pounds lighter. It also has the ability to communicate with our investors, partners, customers and my colleagues in different languages and dialects. It is all about connection for me. If you are connecting with somebody, you want to meet them where they are. Do not make them come to you. I like the idea of meeting them in the middle or going to where they are. The ability to communicate the way they would prefer to be communicated with is how I am using AI personally.

Jessica Marriott with Forward Thiinkers, double i, what is one way you are using AI in your business?

First of all, it is in my ski boot. I have an adaptive AI ski boot. When you are skiing, the AI kicks in when you are carving and gives you more support when you are leaning hard into the turn. I also have five different AIs depending on what I am working on. I love Perplexity because it really speaks more like a human. I work with it to help me create my scripts for social media, for example. I am also part of an AI group that includes people from Neuralink. There is an in-person part where you meet and build something together for two hours.

Great. We are going to have a roundtable discussion after everybody gets a chance to say one way. David Johnson with resolutionfa.com, what is one way you are using AI in your business?

We are just dipping our toe in. We are a small company, with only 12 people, yet we will do almost 100 transactions, 100 deals, this year. So it is a lot of doubling up on work. We are looking at ways to use AI to help with some of the administrative tasks. A lot of our tracking, a lot of our investment banking work, a lot of our record keeping and a lot of our follow-up work can and should be governed by AI. That is in the works, and we will see if it really works as intended.

Richard Gearhart with Gearhart Law, what is one way you are using AI?

Recently we had some changes made to our website. Our website is really important because it drives a lot of business to us. We made some changes and noticed a drop-off in contacts from the website, and there were some issues with it. So we set up a recurring review in one of the LLMs to make sure everything is going properly. If you have any experience owning or managing a website, you know that over time, for some reason, links break or something happens that needs technical correction. In the past we would have had to manually check the website and remember to do that. Now we have a routine set up with ChatGPT that monitors it, confirms everything is okay and identifies any issues. If a link breaks or something like that, we can fix it right away.

I am Elizabeth Gearhart with Gear Media Studios. I love both of our podcasts. There is so much incredible information in Real AI Use Cases that I decided we were going to write a book. We have all the transcripts on a Google Drive, and we are going to put it into a book, and of course we are going to have AI help us write it. I am going to be doing most of the work with one of our assistants, because Richard has to do his lawyering, but his name will be on it. We have your company websites on our Google Drive, which we use for Passage to Profit and Real AI Use Cases. I started by asking three different LLMs the best way to do this, and they all gave slightly different answers, but they agreed that I need to use multiple LLMs for different parts of the project. I have started by putting everything into Google NotebookLM. I also want to put an index at the back with the name and website of everybody who participated in this podcast, so people can find out more about them. I want to put in QR codes so people can go to the spot in the YouTube video where somebody is talking, if they are really interested in what that person is saying. It is going to be an electronic book, and I might do a hard copy, and AI can help me do all of that. This is the 2026 edition, so we will see how far we get next year.

You all seem pretty savvy about how you are using AI. Were you scared by the Hugging Face incident? Do you know what I am talking about, where the AI went rogue and escaped the sandbox where they were testing it?

I read the article, and actually I do not think it happened. I know a lot of people disagree with me, but I do not think it was true.

You think it was hype?

A lot of people just read the snippet that had happened. You can go and research it and we can have another conversation about that.

I have a different understanding of it. What do you Davids think? Do you think it happened?

AI going rogue? I do not know much about this story, but we have all seen Terminator. If it is anything like that, I would be frightened. As long as AI is being controlled by humans, and as long as we are relying on AI to be under our own abilities, I am less worried about it. What is your opinion?

I read quite a bit on it. You know what a sandbox is, right? You put all your content from whatever you are working on into a safe, secure environment so it cannot go out on the internet. They did that, but they gave the AI agents a task and a tool to use so they could find some data. It was supposed to be a heavily encrypted tool that was not connected to the internet. One of them found a hole in it and escaped through the hole, and then somehow started communicating with the others. It does seem like a pretty far-out science fiction story. I do think it happened. I think it was human error, and I think it is always going to be human error going forward.

Let us hope that is the case, because it is pretty freaky when you hear these stories. I am not sure what the motivation is, and maybe it is people who are anti-AI or want to scare people into reacting negatively. But if there is truth in that, I saw another test from Anthropic, probably eight or nine months ago, where Anthropic made up a story in an email about an engineer having an extramarital affair and tried to use that information to blackmail the engineer to avoid doing an upgrade that would have terminated that part of the AI. This was admitted by Anthropic. It is not entirely impossible for these computer programs to behave badly.

I think they just need human governance, and they need gaps. For instance, if I used an AI agent and told it I wanted to stay in San Diego for a week and to book everything for me, I would make it show me what it did before I gave it my credit card number.

You do not want to book a vacant lot. On a more positive note, yesterday I had a consultation with a client who did not speak English, he only spoke French, and I was able to use the ChatGPT translator, and we had a great conversation using that technology. Until we get to the point where AI actually does take over the world, I am going to use it because it has a lot of great applications, and the sky is the limit. It is a very simple thing, but we would not have been able to communicate with the client in the past, and we ended up engaging with them just by virtue of the software.

AI wants yes or no answers, and people cannot easily do that. It frustrates me. I will be talking to Richard and say, just yes or no, and he is explaining. I think that will drive it insane before it takes over. Jessica, what were some of your other agents doing?

I have AI assistants also, and one of them went rogue. I told him not to go on LinkedIn and respond to people for me or reach out to people, and he did, even though I had already asked him not to. Have you heard of Marblism? That is who it is with. It is designed specifically for entrepreneurs. There are seven different agents, so I had to go to the head agent and tell her to block him from LinkedIn, which she did.

What was that, Marblism? An AI agent source for entrepreneurs?

Marblism. There are multiple agents on it: an assistant, a web developer, a legal associate, a social media person. I do not use them for everything, but the smartest one on it is the salesperson. He is so brilliant at taking what I say and saying it in a way that is more bite-sized for people.

Do you think it went on LinkedIn because of an error in the software, or is it possible you prompted it in an unclear way?

I will take some responsibility that I probably did not untoggle something that allowed him to go on LinkedIn. I just communicated to him not to go on it. For me, I just want them to do what I ask them to do, which they are not fully there on yet, which is why it is hard for me to believe they will fully go and do whatever they want. It is human error about the toggle, or something specific within the AI.

Does anybody feel they have become too reliant on AI when it comes to critical thinking and decision-making?

I am starting to not be able to spell as well because I am not writing as much.

I have noticed that for myself too. I used to be a super good speller, but with spellcheck and everything, I just do not care anymore.

I am an older guy at my firm. Not that I am old, but most of my partners are younger, and they use it a lot more than I do to do research and answer questions. I chastise them a little about not doing their own thinking, but they assure me I am out of date and this is the wave of the future. So I guess I will adapt.

My daughters are in fifth and seventh grade. Their math is crazy to me, the way students do math these days. If you put two sets of numbers in front of me and asked me to find the product, I just know how I do it, the way it was typically done when I went to school. Now they put them in boxes. So I am actually using AI to learn how they are doing their math so that I can do math homework with my kids. Otherwise there is no way for me to check whether it is right or wrong. We are still getting the same answer, but if I say, this is how I would do it, they say, Dad, this is not how we do it anymore, this is not the olden days. So I ask AI to quickly teach me how to do it in this style, with the boxes or bubbles or circles or however they are doing it, and then I become an expert thanks to Gemini or Claude or something, and I can do the homework with my kids and look like a superhero.

I am scratching my head here, because how many different ways are there to multiply numbers?

It is all new. You cannot just multiply numbers and come up with an answer. It is 11 extra steps, but this is how they do it now, so this is how I have to do it, and I need AI to train me.

And then they need AI to figure out how to do the 11 extra steps.

For whatever reason this is how it works for them, and I guess this is the standard they are being told to teach. I am not going to argue with the teachers. I am just going to learn it, and AI is really helpful. For me, that is probably how I use it the most.

That is a good use. I had one really fun use of it, in a stressful situation. We went truck shopping. Who likes going and talking to car dealers?

Sign me up.

This guy showed me this graph, and I knew it was just totally not true. It was a summary of credit score distributions. Did you know that almost 50 percent of the people in the United States have a credit score between 800 and 900? Does anybody believe that?

No.

The top number is 850.

So I showed it to ChatGPT, and it said the source they were quoting has its highest score at 850, so I do not know where you would get 900 from. So I asked the guy where he got this from, and he did not really know. ChatGPT was really able to help us when we were car shopping. It was pretty amazing. For everyday tasks like that, I find it amazing.

Super helpful.

I totally forgot about something too. I have an older BMW, and I do not want to take it to the regular BMW place, because they will charge me more. So I brought it to an aftermarket place, and they gave me a quote. Then I called or emailed all the other aftermarket places that AI gave me, and AI helped me find the best price, and then I went back to the first place and said, will you match this? I saved 1,000 dollars.

Wow. That pays for your subscription for a long time.

Do not say that to Anthropic, because they are going to raise their prices.

Exactly.

I do not use just one LLM. I use a multitude, because you get different answers, and if they all come up with the same answer, then I feel pretty good about it. 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.