Did you know I speak Spanish fluently?

Unfortunately, I don’t. But with the help of AI, it’s a virtual reality.

Using a tool called HeyGen, I uploaded a video that features at least 30 seconds of me speaking. It then regenerated the video from English to Spanish.

It didn’t just generate captions and translate them (it didn’t do that at all), it cloned my voice and regenerated the audio. It also visually edited my mouth. The final product was that I look and sound like I’m speaking Spanish — in my voice.

It’s not perfect, but it’s pretty darn good. I could see YouTube, TikTok, and Meta building something like this into apps to make videos more accessible, regardless of the language you speak.

I’d love to hear from native Spanish speakers on the accuracy. What do you think?

The post AI-Powered Video Translation appeared first on Jon Loomer Digital.

This is the biggest issue with my bot right now…

I’ve been testing an AI chatbot for a few months. I’m using Chatbase, and overall it’s been amazing.

It’s trained on more than 150 pages of my website and it’s basically been an extension of me. In most cases, it does a really good job of answering questions close to how I might answer them.

But, here’s something it keeps doing…

Someone will ask it for a list of links to blog posts. The bot will provide five links like it does in the video. Overall, it seems like a helpful reply.

Well… It’s super helpful until the person clicks on any of the links. Why? None of the blog posts actually exist!

It’s so weird and random because it will do a good job of sharing links properly when they naturally come up in an answer. It seems that when people specifically ask for links, that’s when it loses its damn mind and makes stuff up.

Anyone else see this? Got a solution? I’ve tried everything, I swear.

The post Why Does My Chatbot Keep Making Up Links? appeared first on Jon Loomer Digital.

This is the amazing byproduct of having an AI chatbot. In order to make it as valuable as possible, you’ll improve your website in the process…

I’ve been testing a chatbot for the past few weeks. My goal is for it to be an extension of me and my content. So I trained it on close to 150 pages of my website.

But something really interesting has happened…

When the bot refers to something I said that is no longer true, I’m motivated to update it. When it seems clueless about how to respond to something, I create content to train it on.

I’ve written nine different pages related to my advertising philosophy to help the bot understand how I look at ads. I’ve written pages like “Where to Start” because this is a popular question people are asking.

In the past, I simply accepted that things get outdated quickly. That’s just the nature of Facebook ads-related content.

But now, it drives me crazy when the bot says something that is no longer true. I want to immediately correct it, update it, and retrain it.

Looking through questions people have asked the bot has helped me see pain points I haven’t addressed before, and opportunities to provide more clarity on certain topics.

In the process of creating content to improve the performance of my bot, I’m improving my website. It’s honestly kind of cool!

The post The Byproduct of Having an AI Chatbot appeared first on Jon Loomer Digital.

This is what worries me about AI-powered chatbots…

When you create a chatbot, you provide the content to train it on. You can provide files, text, or website URLs. You can have it crawl the entire sitemap of a domain or submit links individually.

The potential issue is that you can submit any content and any links. It doesn’t matter whether you own that content.

You could then build and market a resource on a topic that is trained entirely on content that you do not own. That seems… wrong.

This, of course, is a philosophical debate about ChatGPT and Large Language Models generally. But, whereas the source of a specific answer from ChatGPT may be a bit murkier because it’s derived from a compilation of information, this feels more clear-cut.

When you create the bot, you provide the content to train it on. The source is clear to the app creator and app developer because it was submitted to create that bot.

There are some potential solutions that would fall on the app developer:

1. Require domain verification before submitting links or a sitemap. This seems like a reasonable requirement.

2. Create a tool that scans submitted documents and text for matches on the internet. This could trigger a notification to the content owner or a restriction that would prevent that content from being used for training.

Possibly the easiest solution: Require all bots to link to their trained material.

Ideally, every answer will include a source. But even without that, a link to the full training material would provide some transparency. There’s little argument for hiding this information.

Right now, there’s little preventing someone from building a bot (or another AI-powered tool) that is trained on another website’s content and then marketing it as their own.

Of course, lawsuits will be coming, and the type of transparency that I recommend here could help.

What do you think?

The post The Problem with AI Chatbots appeared first on Jon Loomer Digital.

This is the latest iteration of my AI-powered chatbot.

I’ve experimented with seemingly every possible app, but this is likely the direction I’ll go. This time, I’m using Chatbase.

I’ve trained this bot on more than 140 pages of my website. My hope is that it will be able to behave as an extension of my content.

I’ve created two chatbots. One is trained entirely on Facebook ads-related content. The other is trained on short-form video tips and strategies.

I’ve added the bot that’s trained on short-form video information to my Short-Form Video Blueprint training. Adding this bot to the training can make that bot super focused on that content and give the students of the training access to quick answers.

I see huge opportunities for these bots. I’m planning to add one to my core website to quickly and easily answer visitor questions and direct them to related content.

I plan to open up beta access to my main bot soon. Message me if you’re interested!

The post AI-Powered ChatBot with Chatbase appeared first on Jon Loomer Digital.

I’m testing out an AI chatbot that’s trained on my website using SiteGPT. Still early, but check it out…

I’ve trained it on 25 links in all that include things like…

– About Me
– Posts about my advertising philosophy
– Ads FAQ
– Ads Glossary
– Important and relevant blog posts

It’s still a process of tweaking and refining. I feature some of the highlights here, but it’s definitely not perfect yet. There have been some frustrating moments.

It utilizes ChatGPT, so any issues I’m still having are likely more to do with ChatGPT than with the app itself. It’s also working on GPT-3.5 right now, so I’m sure it will be quite a bit better on GPT-4.

This is the way I’ve envisioned leveraging AI to offer value. A chatbot that is trained on my content and can quickly and easily answer questions and direct visitors to specific content, products, and ways to reach me.

I’m currently beta testing it with my PHC – Elite community, but if all goes well I’ll expand the content it’s trained on and roll it out to a broader group.

Have you tried anything like this?

The post An AI ChatBot Trained on Website Content appeared first on Jon Loomer Digital.

Here’s how I’m thinking about using AI for my business on my website…

Ever since the explosion of ChatGPT and AI tools, I’ve been thinking about how I can take advantage of it. It’s been a bit overwhelming because there are dozens of new and unproven tools every day.

Most businesses are thinking about how it can make them more efficient by automating some of the tasks they’re spending too much time on every day. But I’m looking at it a little differently.

How can I use AI to provide more value to people?

I’ve been thinking specifically about my website, which is a resource of Facebook advertising information. It’s the accumulation of more than 1,000 pages over 12 years, and it’s overwhelming.

What if I could add a chatbot that learns from my content so that visitors to my website could get immediate advice and recommendations? And this information is based on my content and philosophies.

I’ve only started experimenting with a couple of tools, and it really does blow me away. And this is just by giving it my ads FAQ, glossary, and About Me page.

What if I gave it all of my content? It also gives me ideas of cornerstone content that I could create about my philosophies that could be a foundation for the chatbot.

I’m super excited about the possibilities, but I’m still sorting out which app I should use. It’s so overwhelming because anyone can create an app, and it’s so hard to tell what is reputable and what isn’t.

Any ideas?

The post Use an AI ChatBot on Your Website appeared first on Jon Loomer Digital.

This is a crazy example of how ChatGPT sent me business.

Someone sent me a screenshot today of what he experienced on ChatGPT. He asked for recommendations to learn Facebook ads best practices. ChatGPT named me immediately after Facebook Blueprint.

One important point is that not only did ChatGPT recommend me, but it provided a link to my website. This is something many of us have worried about. It’s reassuring.

So, the story continues. This person then clicks on that link and goes to my website. He then signs up for a membership.

He never heard of me before. But because ChatGPT recommended me, he is a new member of my community.

Now, this is a huge benefit. But of course, the question becomes how I can continue to benefit from this — and how you can, too.

A promoted recommendation wouldn’t be any good. So, why was I recommended here? And why have I seen this type of recommendation so many times?

There isn’t really a secret. There’s no hack. It’s a decade-plus of hard work and reputation. There’s no guarantee that ChatGPT will continue to recommend me and send me business.

But I’m going to appreciate it while it does…

The post How ChatGPT Sent Me Business appeared first on Jon Loomer Digital.

I took Google Bard for a spin, and it isn’t great. It made up stuff and it simply wasn’t helpful.

Sure, it was nice that it recommended me as someone who can help you learn about Facebook ads. But after that, it was all downhill.

It made up the name of my company. In fact, I ran several versions of this, and it repeatedly made up a fact about me. It was often related to some made-up book that I wrote.

But the main thing is that it doesn’t let you get any more info. Bard had enough data to recommend me, but it was completely unable to provide any more details.

Possibly worse was that it couldn’t provide links so that I could learn more. Ultimately, this was less helpful than a Google search, and it wasn’t close.

Have you tried out Bard yet? What do you think?

The post Google Bard is Not Ready appeared first on Jon Loomer Digital.