Episode 59: How To Make AI Your Unfair Business Advantage With Gunnar Hood
About Gunnar Hood
Gunnar Hood is the President of WSI-Summit. With a laser focus on B2B and professional services, Gunnar's expertise lies in empowering businesses to enhance their online presence and stimulate growth through innovative digital marketing strategies. A seasoned educator since 2013, he's enlightened thousands with his knowledge on digital marketing, LinkedIn, Sales Navigator, and now delves into the transformative power of Artificial Intelligence for business leaders. In 2012, Gunnar left behind a 25-year tenure in financial services to forge his own path in the digital marketing realm with WSI. His relentless pursuit of excellence hasn't gone unnoticed. Over the past two years, Gunnar's office has emerged as a top performer within the WSI network. He's also been adorned with prestigious web design awards from WMA and earned the title of the top social selling influencer by none other than The LinkedIn Whisperer, Brynne Tillman. Gunnar’s joining Prospecting on Purpose as we discuss 'How To Make AI Your Unfair Business Advantage.”
Some people think that AI will replace them in their jobs, but what it does is it automates and makes jobs easier for them. When you learn to harness the power of Artificial Intelligence, you begin to put your organization or business at an advantage. In this episode, Gunnar Hood, the President of WSI-Summit, shares his insights on the power of Artificial Intelligence and how to make AI your unfair business advantage. He also shares his recipe for a successful prompt and how to use it. Gunnar explores the Human in the Loop, CustomGPT, ChatGPT, Image creation, and more! So, let’s join Gunnar Hood in today’s episode and indulge ourselves with his invaluable insights.
---
Watch the episode here
Listen to the podcast here
How To Make AI Your Unfair Business Advantage With Gunnar Hood
I am thrilled to introduce Gunnar Hood, the president of WSI Summit. With a laser focus on B2B and professional services, Gunnar's expertise lies in empowering businesses to enhance their online presence and stimulate growth through innovative digital marketing strategies. A seasoned educator since 2013, he's enlightened thousands with his knowledge on digital marketing and LinkedIn Sales Navigator, and now delves into the transformative power of artificial intelligence for business leaders. In 2012, Gunnar left behind a 25-year tenure in financial services to forge his path in the digital marketing realm with WSI.
His relentless pursuit of excellence hasn't gone unnoticed. Over the past few years, Gunnar's office has emerged as a top performer within the WSI network. He has also been adorned with prestigious web design awards from WMA and earned the title of Top Social Selling Influencer by none other than the LinkedIn Whisperer, Brynne Tillman. Gunnar is joining us as we discuss how to make AI your unfair business advantage, and I am thrilled to dive into his forward-thinking approach to leveraging these new technologies. I for sure know I need it. Gunnar, welcome to the show.
Thanks so much for having me, Sara. I'm delighted to be here.
I have a bunch of questions for you. I don't want to say talk to me like I'm a baby but a little bit. On a scale of 1 to 10, my AI knowledge is probably around a 2 if that helps you.
Don't feel bad about that. I was teaching a class. We went around the room and asked people to say, "On a scale of 1 to 10, where are you on AI knowledge?" The vast majority said 0 to 2. I even had somebody say double negative zero if you can make that. For some people, it's a daily conversation, but for others, it's still this unknown element. My goal is to have conversations with people and get them a little bit more comfortable with it than where they might be now. If they can experiment with it and find something helpful to them, that then starts the whole flywheel process that they go, "If it can do that, I wonder if it can do this too." We can talk a little bit more about those examples, but I know you want to first start with what it is.
Can we do a basic entry-level definition of what is AI? I'm excited to get into some specific examples because that will help.
One way to look at AI is things that we may have been using in our lives already that use it from a corporate level. If I say it, forgive me if something happens. You use the Alexa app on Amazon or Siri on your iPhone. Those have AI as their backbone. You think about it at the corporate level if your bank has ever called you saying, "Did you use your credit card for this or that?" when you didn't.
There's fraud protection happening in the background. Even if you're applying for a credit card, the whole scoring process that goes on behind the scenes is involved with that. On Amazon, if you're shopping for something, in the reviews and recommendations that come up of people who saw this also liked this, there's an element of artificial intelligence or supercomputing going on behind the scenes.
Those are things we have learned from or used on a regular basis without even realizing it. We're transitioning into a period where AI used to require supercomputing power, machine language models, and things to be able to run. That limited it to large corporations or governments. It's now been disintermediated to the sense that you and I as individuals have access to the power of these tools. It opens up new opportunities for us to do things that we want it to do, not necessarily what others want it to do with us.
That's interesting. Amazon suggesting a purchase is Amazon trying to get us to buy something with their supercomputer AI, but now we have access to it. It only got on my radar when ChatGPT came out.
Why did that happen?
We could get it for free on the internet with a web browser.
There was a study done that said only 14% of adults in America had even heard of ChatGPT, yet by that time, there had been several hundred million users of the product, making it the fastest-adopted technology ever in history. Still, there's a large percentage of American adults who didn't even know what it was.
It gave me goosebumps a little bit to think about how fast that technology was adopted, and people still didn't know about it.
One of the analogies we draw to help people understand what it's capable of is the local library. Flashback even a couple of years ago, whether it was Google or the library, if you had a question, you would go on the internet and look for an answer. If it's Google, you get several results that are presented to you. It's like opening up the books at the library and trying to figure out which one has the best answer for my particular needs.
With AI, the difference is that instead of asking the librarian for the information and her or him pointing you to the books or the section you need to reference, that librarian has now read the entire library and can answer the question for you. With the way we have already advanced with ChatGPT, that librarian has now read every book and article in many cases that's out there and can answer the question directly for you, saving you time and effort from having to figure out what those resources are.
With how we've already advanced with ChatGPT, it can answer the question directly to save you time and effort in figuring out what those resources are.
That's an excellent example. It reminds me a little bit about the difference between coaching versus consulting. Coaching is a lot of pulling it out of you and making sure that you understand almost from within how to do it versus consulting, which is, "Here's how to do it. Here's the answer." Both are valuable, but if you want the shortcut, that's why you hire a consultant.
A lot of people hesitated to do things with computers because they always felt, "I didn't know programming language." You may have been exposed to it in high school, college, or something but unless that was your discipline, it probably went in one ear and out the other. The beauty of AI is that it uses English as its programming language. You don't have to know a different language. You just have to be descriptive in terms of what you want. The beauty is that it's called ChatGPT, not AskGPT for a reason because your answers get better as you have a conversation with the tool.
The beauty of ChatGPT is it gives you answers, and it gets better and better as you have a conversation with the tool.
Somebody called them Chatty, and I thought that was cute, "Chatty says this." I was like, "That's cute. I'm going to go start calling mine Chatty."
If you're an Iron Man fan or something, you call it Jarvis. People nickname a lot of different things.
What's the craziest nickname you've heard?
It doesn't even come to mind because not too many people share their pet name with you.
You know mine. It's Chatty. You talked a little bit about the weird start. When I play around with ChatGPT, I don't like the responses that I get, and I know it's probably because I haven't prompted it in the right way, but that's where I get a little stuck or frustrated. I go to use ChatGPT. It spits out something that doesn't sound like me that I wouldn't use, or it doesn't feel authentic to me, so I don't want to put it out there like me. Can we get into some examples of best practices on how to use ChatGPT or give some examples of how we can start to use it in our roles at our jobs?
That's what's going to help solidify it for a lot of people. If they hear one of these examples and say, "It can do that, and it's that easy," they're willing to try it. The encouragement I give people is, "You cannot break it. Don't be afraid to try it. Put in a prompt and see what it means." We should start with what a prompt is because some people may go, "Is it a question? Is it a statement? What is it?" It is a description. Do you want to describe one of the ones you were testing earlier, and we can debug it?
I wanted to prepare some interview questions for this interview, and because it was an AI conversation, I thought it would be appropriate to try. I put your name in and then got all paranoid that your name wasn't there. Do people worry about sharing information with it? I'm taking this off a tangent, but is there any harm in that?
That's a good safety point to bring up because it isn't 100% secure. Some people may work for corporations that prohibit employees from using it on their computers, and the reason for that is they don't want any of their protected information to end up on the web. In some respects, that's what it is. If a corporation hires somebody like Microsoft, Amazon for Azure, or AWS as their backbone, they can create a firewall version so that the company data is protected, but they still have the ability to reach out and access ChatGPT for its wealth of information without compromising their data.
Be careful about uploading intellectual property into the tool. The prompt that I wrote about preparing for this interview is, "Write a list of ten interview questions. I'm interviewing a digital marketing specialist, Gunnar Hood, from WSI. We're going to discuss how AI can be your competitive advantage at work." Out of the 10, I liked 7 of them.
Did you follow up with that?
I did because I've been playing around with it. I said, "Please remove questions 2, 3, and 9 and expand on questions 5 and 7." Question five was practical examples. That was the response it gave me. The question from ChatGPT was, "Could you provide some practical examples or case studies where the integration of AI had a significant positive impact on companies' digital marketing efforts?" That's not the question I asked you because it's not about digital marketing, but because I included that you were the expert in it, that's why it spit that out.
The only thing that you could have shaped differently in your prompt would be, "While he's an expert in digital marketing, let's ask questions about how it might apply to a small to medium business owner and their operations or their business." Adding that little bit of context will reshape the questions a little bit.
Is there a recipe or a formula for a successful prompt?
Some of it is telling it to assume a role in the first place. Most of mine will start with, "Assume you are a data analyst, a digital marketer, or an SEO expert. You are a CPA or even a lawyer." You're giving it context to say, "This is your role. This is your area of expertise." You want to define what it is you want them to do, and you might say, "You have a client that is in XYZ industry. They are having a problem with this and have come to you for advice around XYZ." Give it more details. The more information you can feed it in terms of details, the better it gets.
The analogy we use here comes from my sales coach. He says, "Generic questions get you generic answers. Specific questions get you specific answers." The way to look at that is you could tell your GPS tool, "I'm in Houston. I want to go to New York." It could take you any way it wanted but if you said, "I want to go to New York but I need to make sure I go through Memphis and Akron or something else," you're giving it much more explicit directions, and you can also say what to avoid, "I don't want to go through Florida. I don't want to take I-10," or things like that. When you give it both the positives and the negatives, it creates guardrails for it to operate within and give you more explicit responses.
That was a very helpful example because looking back on all of my prompting, I do think it's generic. I take it 40% of the way there and then get irritated because the responses are repetitive, or it's not what I was looking for. I like that. Give Chatty a role, assume you're XYZ, and then define what the task is. Let's get into some more specific examples. Specificity is king.
You also don't have to feel that you have to get it 100% on the first try, hence the chat. You liked a couple of responses and said, "Give me more like this and less like that." That's good feedback for it because now it can iterate down. Usually, if you go through 3 to 5 iterations, you will get a pretty good outcome but if you expect it to be the very first one, you're probably going to be disappointed a lot of times.
That's fair. We have the show example. We have the Houston and New York example. Let's get into some more specifics of how people can use it. Start our juices flowing.
In the workshop I gave, one of the examples I talked about was this. Let's say you're a business owner, and reviews are important to your business. You can download your reviews from Google and put them in a spreadsheet. You can then feed that spreadsheet into the paid version of ChatGPT and ask it to analyze that and give you a sentiment analysis. This could be something an intern does at your company but the goal is, "What can we learn from customer feedback to help guide improvements in our business that make for more satisfied customers?"
You can have it assign a waiting to types of reviews, get quotes around it, and so forth. You will be amazed at what it can churn out. There was a person there who told me that they started doing this. They run a division of the Better Business Bureau. They review customer sentiment quite a lot, and they said what used to take them a full day to process data around an industry they can now do in two hours with the help of ChatGPT.
I always give out feedback surveys after my workshops in person or virtual. I have been seeing some themes but I haven't done any type of concrete analysis on what's the biggest impact because I usually ask people, "What is your biggest takeaway?" It's usually 1 of 4 things. Each time I redo it or give it again, I'll emphasize that section more but I never thought to use ChatGPT to summarize it.
One of the things I do is teach LinkedIn. Before I start with a client, I send them a multi-question survey to assess what are the existing skillsets of these people. I then take those results in spreadsheet format and give the instruction sets to ChatGPT to say, "You are a LinkedIn trainer taking on my role. You've collected this information. You're trying to identify where to customize the training to fit the needs of the organization. Identify the skill gaps and knowledge levels and develop a plan of action that you can discuss with the client."
The specific details are in the survey questions and answers. I can attach those to the prompt and say, "Here's the information." It takes a minute or two to digest that, and sometimes it comes back and says, "I'm not sure what's in this column. Tell me what it is. I need to go back and review it again." It will come back, and you can even say, "Show me graphs."
If you're familiar with the Social Selling Index on LinkedIn, you can have it graph a histogram of performance or scores by individuals on the team to say, "What's the distribution here in each one of these categories?" That can say, "This one is on the lower end of the scale, which says there's a huge opportunity for us to focus on that and move that ahead for you." It used to take me hours to compile it and put it into graphs and things like that on my own. It can now do it in twenty minutes for me with a lot of back and forth.
That's interesting because I want to pull that thread a little bit. I read a Forbes article about how white-collar workers who do not embrace AI are going to get left behind, and that intimidates me. That's why we're having this conversation because I want to learn more about it. That is such a great example of how if you don't embrace it, it will take all day. If you do embrace it, it takes twenty minutes, and then you have more time to do whatever your roles and responsibilities are. I'm assuming that's what that means from that Forbes article.
There have been a lot of articles published from the New York Times to Forbes and others, and some people will walk away thinking, "This is going to eliminate jobs." That's not what it does because it's very task-specific. What it does is it gets good at tasks. The fact is that if you can automate a percentage of your tasks, you don't need as many people to do all of those tasks but the ones that you do need are the ones who understand how to use AI. If somebody is resistant to learning AI in an organization, they will be at greater risk of losing their position because of that to somebody willing to do it.
There's a quote that I share with people a lot. I'm probably paraphrasing it but the quote is, "Success will come to those who learn new technology the fastest rather than old technology the best." This may date me but some of the earliest word processors that existed out there were things like WordPerfect. This was before Microsoft Word. There's not a need for a lot of WordPerfect experts out there, and there's less of a need for Microsoft Word experts because there are other tools you can use. If you are willing to invest some time and learn these new technologies, not only will you get benefit from it but it's going to insulate you against the competition that is slower to adopt those technologies.
Let's talk about that a little bit more in terms of differentiating ourselves because I liked the example of summarizing customer reviews. Can you give us a couple of other examples of how we can start to pull it into our day-to-day activities?
Let's say you're sitting in a meeting with one of your clients. They happen to make a reference to an article that they read or a book that they thought was good, and you're not familiar with it. You can ask ChatGPT to go out there and summarize the book on a chapter-by-chapter basis or summarize the article for you so you can get a sense of what's in there. You can even take that a step further.
I do this with YouTube. You can watch a lot of YouTube hours of videos and get disappointed at the end of it going, "That's not what I wanted," but you can use a Chrome extension with YouTube and ask ChatGPT to give you a fifteen-bullet point summary of what's in that video, and it completes it in a minute or two. You can say, "That looks like that's the right video. That's not the right video. Let me keep looking for another one." It can be a huge time saver and make you happier with the content that you do want to consume versus disappointed by it.
That was a great example too. I'm thinking about some of the areas where I got stuck and still do like writing emails, for example. Sometimes emails can take so long. If you have a delicate email to send, can Chatty help us? The stakes are high. I call them crucial conversations.
I've read those books.
I haven't tried it yet but I'm imagining it could do some of that too.
Here's what's cool about it too. If you've read the book Crucial Conversations, so has ChatGPT. You can instruct it to apply the methodology used in Crucial Conversations to help you frame that email.
What would a prompt like that look like?
I assume that you are an expert in the topic of Crucial Conversations, helping a client draft a sensitive email on XYZ topic. You could take the first draft and say, "Here's the first draft of the email. Please apply the Crucial Conversations methodology and improve on this." That's step one, and then you look at it when it comes back. The risk with anything that it creates that you send to somebody else is that it can sound like AI. There are some pretty classic indicators in there. One of our clients sent us a blog article to read. They didn't mention it was written on AI but the final paragraph starts by saying, "In conclusion," which is a classic telltale sign of AI. We're like, "Did you use AI?" "I did."
We describe a concept called the human in the loop, which means you don't want to trust implicitly that it's going to get it right but you want to add your voice and your tonality. Brynne Tillman who we mentioned earlier and Bob Woods, her partner, often talk about reading it out loud. Does it sound like you or not? If it doesn't, then tweak it to make it sound like you. You may like what it says but you still want to tweak it to be your voice.
That's a great easy tip too. Bob is going to be on the show. We didn't spend a lot of time on AI because we knew we were talking to you. I like the idea of reading out loud. For my season two premiere, I had asked ChatGPT to write an intro paragraph to see. I was playing around, and it was, "Prospecting on purpose where we ignite the flames of your personal and professional growth." I was like, "I don't think I would ever say that." I was trying to find a way to make it a joke. I'm like, "We're over here igniting your professional growth." The human in the loop makes a lot of sense to me.
That's another area where humans will continue to play an important role regardless of the adoption of AI in the future because if you have a level of expertise, whether it's a creative genius in a particular discipline or something else that might be difficult for AI to replicate, you are going to pick up on what we call the hallucinations that exist. They have gone down as it gets smarter and more people give it feedback but hallucinations are where it can make stuff up. There are prompts you can use to make it fact-check itself these days.
You want to know your content. An example that became very well-known a while back was a lawyer in New York who was preparing for a trial, and he used ChatGPT to pull up a bunch of what he felt were relevant cases that he could use as examples in the courtroom of why his client should get the outcome he wanted. It turns out every single one of them was made up. They were not real. They were not factual. The judge understood that and tossed the guy out of court pretty hard saying, "Don't ever do that again." The downfall was that he didn't fact-check it himself. He wasn't a human in the loop. He took it at face value.
I understand the concept of how the tool is always learning and self-improving. I get that pretty clear but why would that have happened? Where would it get the made-up stuff? It made up cases.
It wants to give you answers.
That was cute. It's like your dog wanting to please you.
Before there was ever GPS and things like that, there were always these comedians talking, "Ask a guy for directions. You will get an answer but it may not be the right one."
That's a great example. It's the guy at the bar at the gas station.
"Go down there far piece, turn left there, and so forth." They want to be helpful.
Nobody wants to admit they don't know either. That's probably part of it too.
How much of that plays into it? I'm not 100% sure but in the year that I've been using it, I've seen it get much better. I also recognize my prompting has gotten a lot better. You can ask it and say, "Go back with what you gave me and fact-check this. Show me your references." Early on, I would get references. I clicked on the links, and they would be dead-end links. That didn't work. It would go, "Let me try again." It will say those words to you, try again, and get some stuff. It's a great thing.
There's another prompt that I wanted to share with you too that can be very helpful for people starting and going, "I don't know what to ask." We've got an eBook out. One of the prompts here is called What Other Details Prompt. I love this one because you can give it certain information but if you go, "I don't know what to ask it to give me," you can have it prompt you to ask you questions based on that. The prompt is simply to enter whatever prompt you want and then say, "What other details do I need to provide ChatGPT to ensure maximum success?"
That's helpful because, in my example of the interview, I could have asked what other details were in it. Maybe it would have said, "Who is your audience?" I'll try it and see what it says.
You can ask it more than once as well. If it doesn't ask you a lot of questions, or you're not sure, going, "That's interesting," what else should you be thinking about or would help you? It is a great way to get it to train you how to use it at the same time.
We talk about sales and prospecting a lot on this show. It is a sales-lensed show. How can sales professionals use ChatGPT to help them in their professional development differentiate themselves from the competition? What are some tools we can use ChatGPT for in a sales role?
We know that some of the highest-performing sales professionals out there have a system. It could be Sandler, Miller Heiman, or Challenger Sale. Whatever the system is helps them in their process, and they will coach themselves or coach others to that process. One of the great things is that if you can take that process and put it into text, "This is what the process is. This is what a perfect call sounds like. This is what five sounds like," and things like that, you can then upload that information into ChatGPT to say, "Here is the sales process."
It's harder to do in person but if you're doing Zoom calls, you can use Otter or things like that and get the transcription. You can then upload that in the same conversation and say, "You are the sales coach evaluating the performance against the metrics in this tool. Review the call, score each component, and provide a summary of what gaps there are and what opportunities there are."
The cool thing is you may be a solopreneur who embraces a particular sales model but you don't have a paid coach or somebody to do that for you. If you start doing these things, you've got an AI coach that can help you get better. Even if you listen to your call, you are likely going to miss things, going, "Did I do that? Did I not do that? It sounded good enough to me." ChatGPT is going to be agnostic about that and score you the way that it has been asked to score you. It's going to be much more objective than subjective about it.
I have six questions from that. When we were talking about the prompting, there seemed to be a formula. Give it the context. In the example you gave, the context would be whatever sales methodology you're using. Context first, and then define what you want. The request is, "Review this transcript of the sales call and give me a rating or feedback on areas of opportunity."
Evaluate the performance against the sales model of this. You can make it that simple and tweak it afterward.
That's cool because most people, if they have a growth mindset, want to get better, and they're open to feedback even though it can be tough but it's almost like this is not your boss giving you feedback or the customer giving you feedback. It's not a real-life thing. It's an objective tool that we can start using, and it's not coming from a place of ego or objective viewpoints. It's more straightforward, and there's no emotion involved in ChatGPT's response.
If you take it a step further, oftentimes if you're in human interaction or a coaching interaction, there's a hesitancy to say certain things for fear, "I don't want to hurt their feelings," or they don't want to ask because they don't want to make it sound like they did something bad. There was a book written by a New York Times author several years ago that was titled Everybody Lies. The title went longer but what he said was you can distill the truth by looking at Google searches because people don't lie in their searches the way that they lie in surveys. If AI is here as a tool, people are likely going to be more honest and candid in what they want or the feedback that they give than if they were dealing with a human.
You can distill the truth by looking at Google searches because people don't lie in their searches with how they lie in surveys.
I've never thought about that. That makes a lot of sense because people don't want to let you down even in sales. I would have customers or prospects promise me something. I used to sell fire features, and they would say, "We have a fire in this project." I would follow up, and I don't think they needed that or wanted it. They didn't want to let me down because they liked me. You wasted everyone's time because you're trying to be nice. I understand it because we all do it but I like the idea of viewing it as an objective coach and objective resource to help make us all better in our sales efforts.
The other part about it too is it can review far more calls than a human can. In my early days in financial services, I was a coach and a trainer. We would download calls from reel-to-reel tapes, listen to those, and then coach people on them but the number of calls we could listen to compared to the universe was difficult. With ChatGPT, you can do a much larger number and then get trend analysis around it too, going, "Is this person improving or not based on the feedback?"
I wonder if it breaks down, "Sara talked about herself for five minutes." I wonder if I could give you that.
Some of the tools like Otter.ai exist out there. One is called Sybill, and I forget what another one is. They have some of those things built into them as they evaluate the percentage of talk time and other things. Some are more geared toward the sales environment or customer service.
That's another thing. We know ChatGPT. You mentioned a Chrome extension to summarize YouTube videos. I had something I had to do over the holiday break. I have my workshop. It's done. It looks cute. It's my branding and my slides. I'm presenting at an in-person workshop, and they needed me to use their PowerPoint template. My presentation is long and fun. There are animations.
Transferring it from Canva to PowerPoint took me two Harry Potter movies. I watched two Harry Potter movies in the background. That's five hours of my time. Luckily, it was over Christmas, and I wasn't doing much. It wasn't hard but I kept thinking, "This is so frustrating. Why can't AI do this for me?" I'm guessing it could. I wouldn't know where to start exploring that. We will use this as our example. Any task that we think is taking up so much of our time, what's our first step to try to see if AI could solve it for us or do it for us?
Start with the highest level. Ask it if it knows of any resources available that could help you do that, "What would be the steps?" My presentation is in Canva. I need it to match the PowerPoint deck provided by the client. Are there tools available that can help me do this quickly? There isn't a translator because I've looked. There's still some squishiness even with AI in Canva but one of the things you could do is take your client presentation deck and load it as a branded presentation in Canva so that now, you can apply that branded model to yours.
If you use Canva's fonts, they likely aren't going to translate to your clients but it can start to understand, going, "If you want to change this to this font, do you want to change it every time it occurs in here?" It can shortchange it a little bit. Ask AI, "Are there tools out there?" There are hundreds of tools being created daily to do this. That might be very helpful and save you one Harry Potter episode.
Luckily, it was number five, and that one is boring. I never would have thought to ask ChatGPT, "How do I do this?" That's a great first step, "How do I do this? What other details do I need to provide you?" That's clever. I have another thing that could be an interesting thing for us to discuss. I have a proposal that's due to a prospect, and they want some custom offerings that I could do. I haven't put pen and paper and priced it out before. I'm imagining proposals would be something. What other things can we start looking at ChatGPT to help expedite and collapse time for us?
It is exciting because it was in early November 2023 that they announced the most recent release of ChatGPT, which included something called custom GPTs. The best analogy I can give you for this is whether you've got an Android or an iPhone, there's a store out there that sells apps that do specific tasks. You buy that app, or there's a free app that lets you do something specifically like workouts and anything else. Custom GPTs are like that in the sense that you now can build the app. If you are intimidated about creating your app for iPhone or something because you have to know a certain programming language, that's now English. ChatGPT gives you the framework with the paid version to create these custom apps or GPTs as they call them.
It saves you time and effort because you can preload in the assumptions, the guardrails, and other things, "Assume you are the head of your company that does this." You give it the guardrails, and the client then would become the exception to that, going, "Apply all of this knowledge to this situation. We need to create a proposal that includes our basic pricing framework plus unique elements for X, Y, and Z. Put this in our standard format for proposals and include this but leave this out."
There was a digital marketer who attended my session and sent me an eight-minute video showing exactly how she did this. She was like, "This is going to save me so much time going forward," because now it knows her pricing structure, her rules, and everything else. She can say, "Here's the client's name. This is what we talked about that we're going to do. Put this in a proposal." She says, "I used to spend an inordinate amount of time overcomplicating this process, and it has simplified it to a couple of minutes."
I'm going to try it and then send you a video of my efforts too. I appreciate all of your analogies because that's helping me make the pathways in my brain. I like the idea of how your phone has different apps for different reasons. If I'm thinking about a custom GPT for my business, for example, I could have one custom GPT or an app. We're using those words interchangeably. One could be for solo episodes, interview questions, or proposals.
A lot of my content is adapted for the customer. It's customized, which is quite time-consuming but you want the examples to be customized to your audience. Maybe that would be another custom GPT. I'm trying to understand it for my business and how I would use it. Did I summarize that correctly? If I have a proposal, I go to my proposal custom GPT.
Here's another example. A guy I know who also is in the marketing space has a client in the funeral home business. One of the things that they do repetitively is write obituaries. He asked them, "How long does it take you to write one of these?" They go, "It could take 5 to 7 hours to do that." What they did is they took a number of the obituaries that they had written to understand the empathy, the tonality, and the type of topics that are usually discussed in it.
They loaded those, created a custom GPT, and then said, "It's going to ask you certain questions when you have a new one to write." It would go through the process, "Tell us about the person, family members, survivors left behind, and things like that." It would output a new obituary, and the response from the client was, "This sounds exactly like we wrote it but it took 3 minutes or 2 minutes, and that would have taken us 5 hours."
I'm thinking about all of the different ways I can apply it. I think about the things that take up so much of my time that are preventing me from doing the next level in my business, not writing obituaries but things like that, emails, and content producing. LinkedIn posts can take hours to write. I sent out an email, and somebody said, "This was so well-written." I said, "It took me two hours, and we used ChatGPT," but if I have enough of an arsenal or a bank of data, now it sounds like I can make a custom GPT, and it will start to understand my voice and my style.
There's a future in selling custom GPTs as well. The question is, "If it can do that, why can't it copy somebody else's?" The magic is in the database. It's your data, your phrasing, and things like that. If that's protected and not public, then that's what becomes of value to people. It's how you did it, not what it did.
That's where I get a little nervous about uploading things to ChatGPT because it's my intellectual property and my method, and that's what I want people to pay me for but if I'm giving it away for free on the internet in nine different spots, am I cutting myself off? I don't want to have that mindset. That's a little bit of a scarcity mindset but it's there. I like the idea of having the protected custom GPT that's my property.
It could be a great question to ask ChatGPT these days too, "I want to create it but I'm concerned about this. What steps can I take to protect my data?"
The moral of the story is to ask ChatGPT if you don't know where to start. I've been going to YouTube videos, and I'm like, "I'm so dumb. I should have been going straight to the source."
Here's a fun example. One of the things they also released was the ability for it to interpret images. I took my phone to the fridge, opened up the fridge doors, took a picture of the inside of my fridge and the contents, and said, "Chat, I want to cook something with chicken tonight based on the contents of my fridge. What would you recommend?" It gave me two recipe options. It says, "I see you've got honey and mustard. You could do a honey mustard glaze with olive oil and saute the chicken in that. I see you also have a lot of vegetables. You could do a chicken stir fry as an option." If you took those, you could say, "Give me a recipe for those," if you wanted to.
It reminds me of your earlier example about the librarian because I would do the same thing with Pinterest. I would go to my Pinterest app and say, "I have these ingredients in the fridge." It would come up with recipes but you still have to click and read through all the ads and the blog post about the history of chicken from all of these different blog readers to get to the recipe. I am going to try that.
It's pretty interesting what it will come up with. If you don't like it, say, "Try again." Hit the regenerate button.
Knowing you, I have been trying to use a lot of nice manners when I talk to ChatGPT. How do you feel about please and thank yous?
I may be stupid on this point but I've heard different opinions on it, and one is that it's being trained with everything that we do. Does it hurt to say please and thank you and hope that maybe that's part of its training? Others have said it doesn't care. You can be matter-of-fact with it and command it. I find myself doing that because it also responds with, "I'm sorry," and things like that. It seems a little bit more human if you want to look at it that way.
It's always learning. We want it to learn nice manners before our AI overlords take over our lives but until they do, we will stay in the human in the loop driver's seat. I'm going to ask you two more questions. You did a nice job. I'm way more energized and excited about using it than reserved and anxious as I felt like I was at the beginning of the interview. Leave last-minute advice for people who are resistant or nervous about it. What will you leave us with in terms of how to get over some of those fears?
One is to listen to shows like this where people are translating it into a tool that anybody can use. It's not just for corporate use. Anybody can do it. Think of the simplest thing that you want to solve for that you go, "I don't feel like going to YouTube, digging up all this stuff, or going through fifteen questions on Google." Try it even in the free version of ChatGPT. In the free version, the data set is now current to April 2023. There's still some lag to it but if you don't need it to be the most recent minute information, then it's going to have a pretty robust database that you can have faith in.
Try it. You will be surprised. Maybe you're planning a trip. Ask it to be your travel advisor and give you some ideas. How many times have you gone to TripAdvisor and other places and had to sift through so many different things? I can say, "Be my travel advisor. I want to go to this location. I want to do hiking, this, and that. Give me a three-day itinerary that fits this budget," and see what it comes up with.
That's a good takeaway. Anytime you go to YouTube or Google, start going to ChatGPT first and see what it comes up with.
Those tools are starting to adopt it too. Bing now has AI built in. Google is now starting to incorporate AI, and they have another tool called Bard. That's free. You can use it, and it's good for some things. It's like choosing a car. The difference in AI models out there is like choosing car versions. They all have different options and feature sets. Some are better at one thing than others. That's the same thing. You don't have to know what all those are to use them. Test them and say, "This is the answer I got on ChatGPT. Let me try it on Bard and see what the difference is."
The difference in AI models is like choosing car versions. They all have different options and features. Some are better with one thing than others.
That's a good idea.
They reference different things. It's still the universe of the internet but Google has a strength in what it did. ChatGPT got sued alongside Microsoft by the New York Times because the New York Times is alleging that they used their database of information to train ChatGPT, and that's a copyright infringement. That may be impacting some of the information we see there in the near future.
I saw that. I know we're wrapping up, and now we're opening more cans of worms. I went to a conference in May 2023, and they were talking about human-made versus AI-made and what you can copyright. If it's something AI-generated, it's a lot more difficult. I don't think legally you can copyright it. It's going to be a learning curve as we wade through all of these unknown situations because none of us have ever been through it before.
You brought up a great example we hadn't touched on yet, and that is image creation. You're writing an article. Maybe it's for a client. You're doing something specific. How much time have you spent sifting through stock image photography and trying to find the image that fits the scenario?
I tried to find a picture of a winking bird the other day, and I didn't like any of the graphics out there. I didn't even think about AI. I needed a picture of a bird winking.
You can type that into ChatGPT and say, "Create an image of a bird winking." You could say, "Make it a toucan or a parakeet," or whatever you want. It will give you a version, and you can iterate on that. There's a lot of image-specific generation AI tools out there. Midjourney is one. DaVinci is another. ChatGPT has it. Apparently, Google released something. You can now type that into the Google prompt and have it generate something. It's a way that you can combine creativity in images that were extraordinarily difficult to do in the real world because how often would this lion interact with this human in a car driving down Fifth Avenue or something like that? You can make that happen with ChatGPT.
There are a lot of image-specific generation AI tools out there. ChatGPT will give you a version that you can iterate on.
I needed a picture of a winking bird. I've spent so much time, and I didn't find it. I gave up on the plan. Honestly, I'm so excited. I appreciate you taking the time to talk to us and talking to us like we're kindergartners. I'm going to say that for myself because this is intimidating but now, I have six different things I want to try as soon as we wrap up. How can people find you and learn more about you, your resources, and your knowledge sharing?
I appreciate the opportunity. There are a couple of ways. From a social media perspective, my primary channel is LinkedIn. There's only one of me on there but rather than have to type in LinkedIn's way to get there, you can type LinkedGunnar.com. It will take you to my profile. That's the easy way. The second thing is I have a digital marketing agency. It's called the WSI Summit. That's WSI-Summit.com.
To make it easy, we have put some AI resources together on a page. You can go to GunnarHood.com/AI, and we will constantly be updating that. If you visit it, you can come back a month from now, and there are going to be different resources there because we continue to evolve. One of those is an eBook that's about 90 pages long that talks about AI, what it is, and how to use it. It includes tools and prompts based on the industry or what you want to accomplish. It can be your starter guide on how to do this.
I appreciate you keeping that updated because that seems to be the thing. Once you learn it, it's still changing every day or every minute.
The biggest risk for business owners is going, "I see the value of this but I'm hesitant to do it because what if I choose the wrong thing, and it's obsolete tomorrow?" One of the things that we also have that you will find on this page is the AI Readiness survey where you can get a score for how ready you are for AI. There may be things you already have that you're not aware of that are in there, and it will help give you some guidance, "Here's the right next step for you."
Thank you so much, Gunnar. I learned so much. I feel excited. I'm going to circle back with you after my proposal and after I make my winking bird graphic. You're going to get a report back from me.
I look forward to it.
Thank you so much for joining. I am excited about the new evolution we're going into or stepping into.
Thanks so much for having me, Sara. I hope your audience gets a few good tips out of it.
I'm sure they will. Thank you.
