Summarize with AI
This transcript has been lightly edited for clarity.
Introduction and Pulse Check
Gina Ramsey: Good morning, everybody. My name is Gina Ramsey, I’m the president of Pink Dog Digital, and I’m here with two of my amazing teammates, Coral and Cory. We’re going to talk about AI today. I’d hope all of you will jump in and ask questions as we go, rather than us just being talking heads up here. We’ve got a lot of information, and we may end up doing a part two if we don’t get through it all, because there’s a ton.
So let’s start with a quick pulse check: raise your hand if you’ve tried ChatGPT or another AI tool. Oh boy, lots of people!
And the second part: how many of you have cringed at what it produced?
Coral Norman: Yep, everybody’s used AI, but some of the results haven’t been great.
G.R.: So we’ve all tried it, we’ve all had bad experiences, and hopefully today we’ll show you how to get good results and avoid the bad ones. But before we dive in… Coral, Cory, can each of you share what this webinar isn’t, and what it is, so people know what to expect today?
What This Webinar Is (and Isn’t)
C.N.: Sure. First, what it isn’t is a know-all about AI: it’s not a tell-all. AI is big, it’s vast, we’re in the Wild West of it, and things are still very new. So we’re going to share tips and tricks, but this by no means is going to cover everything you’ll ever need to know about AI.
The other thing it isn’t is a catch-all marketing plan for using AI. This is how you supplement your marketing with AI while continuing on with a quality marketing plan, whether that’s with us or on your own. It’s a supplement, that’s all it is. So the main thing we’re going to cover is how you can use AI to supplement your marketing properly, and make sure it’s working for you instead of spitting out absolute trash, which happens more often than not if you don’t have much experience with it.
Cory Wasielczyk: And we’re also going to be realistic about how you can use AI, because if you’re in the space and have some knowledge of the major players, it’s easy to buy into the hype that this is some major, revolutionary way of changing how we interact with the world. It definitely has a lot of potential, but you have to learn to use the tools responsibly and understand exactly what they’re capable of, and what they’re not. So we’re going to try to clarify some of the buzz around the tools and bring it down to earth a bit.
G.R.: Brilliant. And as Coral mentioned, please feel free to ask questions in the chat or the Q&A, whichever you prefer: we’ll answer along the way. This is a Q&A format: I’ll ask some questions, and Coral and Corey will jump in to answer. If you have questions, please pop them in any time. And if we can’t answer right away, we’ll follow up by email. So please jump in with what you’ve got. And I reserve the right to jump in too, because I can’t not.
Why AI Isn’t an Instant Marketing Strategy
G.R.: Coral, a lot of small businesses – and Cory just alluded to this – think AI is the magic button: they can get rid of their marketing person, their administrative assistant, their accountant, all of it. Why isn’t AI an instant marketing strategy in our space?
C.N.: A couple of reasons. First, AI is only as good as the person prompting it: it only gives you what you ask for. So it’s important to know what questions to ask, and to have a strong base knowledge of what you’re trying to get out of it.
The other reason it’s not going to replace everybody is that it takes time to use. Yes, it’ll save you some time, especially on marketing work, but not all the time. It requires time, energy, and knowledge. And whatever it gives you is never going to be your final product. If it is, you’ve got a problem, which we’ll get into later. AI can give you ideas, a starting place, maybe help polish something, but it’s never going to give you exactly what you need.
C.W.: I’ve found this especially true with photos, like Gina mentioned earlier: with the six fingers or the weird hands and graphics AI has produced over the years. People can spot AI now. A lot of people will look at a picture and know something’s just not quite right.
So it’s important to still have a team with real strategy behind it, or your own internal marketing person: someone with the expertise to catch it when AI gets something wrong and fix it, and to make sure the right strategy is behind it. This isn’t an instant strategy. You don’t just ask a question and get a full marketing plan with all four Ps covered and deliverables for your social media.
Training AI vs. Prompting AI
G.R.: Talk a little about training, prompting, and supervision: what it takes to produce useful information. Maybe give some small definitions, because I don’t think everyone in the room knows what training and prompting even mean, let alone what a hallucination is. Let’s cover some basic vocabulary.
C.N.: This is one of my favorite things about AI, so I’ll talk about training, and I’ll leave the prompting piece to Cory, since he’s really good at prompting. My favorite thing to do is train an AI. I liken it to an assistant. It’s a marketing assistant, not your marketing director – somebody who comes in and helps with the smaller pieces you need done.
Training is a big part of that. You put in a prompt, and when it spits something out, a lot of people make the mistake of stopping right there. They pull it, maybe make edits on their own, maybe stick it directly into social media. But the problem is, if you make those edits, you should feed them back: “Hey, this is what you gave me based on what I asked for, this is what I was looking for, here’s how I edited it: next time, can you emulate this?”
I think the big thing people miss is not interacting with the AI. Just telling it what they want once, and then doing whatever they do outside of it without feeding that interaction back. It’s important to go back and tell it what to do and what not to do based on what it gives you: tone, information about your business, exactly how you like things written, not using em dashes every five words, that kind of thing.
Writing an Effective Prompt
C.W.: When we talk about a prompt, we mean the actual text you put into the LLM or AI tool to make the request.
The best prompts do a few things. First, you want to give it a persona, for lack of a better term: tell it what it is. For example, if you’re writing a prompt about an advertising campaign, you’d say, “You are an advertising expert, and you know everything there is to know about pay-per-click advertising.” That focuses its knowledge and gives it a basic framework to follow.
Second, you want to give it information about the query. Continuing the example, if you’re asking for A/B testing ideas for pay-per-click headlines, you’d give it a list, like “here are the best headlines for pay-per-click advertising, do the following: one, two, three, four.”
Third, you give it the actual request: “Please give me five headlines I can use to A/B test against this headline.” You can also tell it the deliverable format you want: “format it as an unordered list, and give me reasons why we should try each headline.”
So it’s more than just, “Hey, can you give me some headlines for a pay-per-click campaign?” The more information you give the tool, the better its output is going to be. Context, depth of information, background… it’s not a one-liner. It could be a page or multiple paragraphs to get the right output.
C.N.: Cory shared a prompt with me recently that was basically to make AI not talk like AI. Do you remember how many words it was?
C.W.: It was 1,200 words.
C.N.: And most of it was just a complete list of every AI writing cliché you could identify, saying, “Don’t do that.”
C.W.: We actually got a tweet from Sam Altman… I think it was last week, where he said, “I can report that if you tell ChatGPT not to use em dashes, it won’t use em dashes.” And it’s like, well, why wasn’t it following the instruction beforehand? It doesn’t make sense, but okay.
Where AI Gets It Wrong: Hallucinations
G.R.: That segues nicely into the next question: what kinds of tasks or decisions is AI not good at? Do you have examples of AI gone wrong that you can share with the room? This is a fun one, there’s been a bunch.
C.N.: One thing we want to talk about is AI hallucinations. AI is trained to give you an answer no matter what. I saw one yesterday: someone asked why adding crushed-up porcelain to a baby’s bottle helps support their nutrition needs, and the AI answered as if it were true, explaining how it would help. Clearly, you should not be adding porcelain to a baby’s bottle, there’s no question about that. It’s trained to give you an answer regardless, so you have to be very careful about what it tells you, and occasionally do your own outside research to verify.
Some other examples I saw this morning: in a New York federal court filing, a lawyer used ChatGPT to cite cases, and ChatGPT made up cases because it couldn’t find real ones. So the lawyer ended up citing false legal cases in federal court and got called out for it. Even when you’re using ChatGPT, you have to double-check what it gives you to make sure it’s real and verifiable, because if it’s not, you’ll run into brand trust issues. People notice when you’re not telling the truth, and they’ll suss out that it’s AI, and they don’t like that.
CW: My favorite example was when Gemini’s AI overviews were first rolling out last year, and there were screenshots of it recommending you add glue to pizza sauce or cheese to make it gummier. That was my favorite!
Going back to what it’s bad at: I’d say anything creative-focused. If you ask it to do a creative task and take what it gives you at face value, it’s going to be super generic, bland, and boring. No personality, no real human creativity. And as Coral mentioned, it’ll flat-out make things up if it thinks it needs to in order to satisfy the prompt: statistics, reports, court cases, whatever helps make a convincing case.
Audience Questions on Training AI
G.R.: I have a couple of audience questions, and I can already see we’re going to skip some sections since we’re going to get into the weeds here. Janet asked: “I don’t have a paid account with ChatGPT, but it still knows who I am and what I do. Do I need to give it a comprehensive description of my business every time to get the best response?” I think this is part of the training Coral was talking about.
C.W.: If you have a free account, you can go into Settings and give it general information about yourself for training purposes. If you’re using it as a standalone without logging in, you’d need to give that information every time.
The easiest way to handle it is through a custom GPT, which requires a paid account, but gives you a lot more flexibility in how you train it. So I’d check the settings and give it whatever information you want there.
C.N.: I don’t think we’ve covered custom GPTs today, but we should definitely do something on that in the future. I do love a good custom GPT.
G.R.: That’s why I said this is intro-level, and we’ll probably do more on AI over the next several months.
Amanda asked: “Can you just ask AI if it made up the information it provided?”
C.N.: I’ve never tried that. We always just say to verify: any time AI pulls a statistic, we go into Google to see if we can find the study it’s referencing. I suppose you could ask it for a reference and see if it gives you one.
C.W.: If you use GPT with research mode or thinking mode, it’ll provide sources for what it’s pulling from. But you always want to check those too, to make sure it’s not misconstruing what the article it’s citing actually says.
Practical Marketing Tasks AI Handles Well
G.R.: Let’s shift into practical tasks AI can help with, specifically related to marketing. What pairs well with AI for small teams?
C.W.: Any analytical task. It’s pretty good at that if you give it the data and ask it to extract insights. Some of the low-hanging fruit of marketing it’s fairly good at too: creating an outline for a piece of content, repurposing published content into different formats, keyword research, competitor research and insights, taking a long piece of text – say, 500 to 1,500 words, or a brochure – and extracting the key points.
Also, any kind of brainstorming: generating content ideas, that kind of thing. Of course, you always want to run it by a human to judge whether it’s fine as-is or needs adjusting so it doesn’t sound bad.
Repurposing Existing Content
C.N.: You covered a lot of mine. Repurposing is a big one, especially for marketers working with content they already have. For example, we’ll take the transcript from a webinar like this, feed it into ChatGPT, and say, “We did a webinar on this: give me five different ways I could use this transcript to create more collateral around small business AI tips.” It’ll come back with ideas: a blog article, a graphic for social media, and so on. You take those and edit them down, and go back and train it further on what you used and how. It gives you the ability to reuse something you already have in ways you might not have thought of, which makes life a lot easier, especially on the agency side, working with a client for eight or nine years doing their social media.
Sometimes we’ll say, “Here are all the social posts I’ve done, here’s the client’s information: help me come up with something a little different and unique.” It may not have the human touch right away, but it gives you ideas you can then add that human touch to.
C.W.: I heard about a webinar recently where they took the actual video of a session, fed it into AI, and asked it to create vignettes: each question becoming its own small video with a transcript and a social post headline. So there’s a lot you can do with existing content.
Content Creation: Ideas, Outlines, and the Human Touch
G.R.: For content creators: how can AI help with outlines, ideas, and drafting without replacing the human? And where have people gotten themselves in trouble with content creation?
Cory, as a side note, can you talk about the case study you did comparing AI-only content creation versus human-touched content?
C.W.: Let me take the first part: ideas, outlines, drafting, but not replacing the human. If you give an AI tool a list of topics or general areas you want to cover and ask it for blog topics or content ideas, you can harness the power of how it actually operates. Think of it this way: if you do a Google search, Google gives you a list of results. AI is like taking that search and doing twenty or thirty other related searches all at once, scouring all that information, and giving you an answer drawn from it. You can use that to your advantage for topic ideas. It’s called a “fan-out query,” which is the technical term for taking one search and running related searches off of it to generate topic ideas.
As for the human element, you have to decide what’s going to work for your business: some businesses have areas they don’t want to talk about, so you always want to double-check that the AI’s output is good for your brand.
Where People Get in Trouble with AI Content
C.W.: Think about it this way: AI relies on training data, whether from other websites or its initial training set; and it can only give you content based on that. If you publish that without human review, it’s going to sound exactly like what anyone else with a similar prompt would get, so you’ll run into duplicate content issues. There’s even the potential – and this is somewhat theoretical, though there are court cases – around publishers suing AI companies for using copyrighted material. Every time you ask ChatGPT to generate content, you’re running the risk of copyright infringement unless you put it through a manual review process.
C.N.: The Chicago Sun-Times and the Philadelphia Inquirer both published fake summer reading lists earlier this year because they asked for a list of books for kids or adults, and the books weren’t real – because they skipped that manual review. That’s how you get yourself in trouble.
Another way is brands relying on AI to model their clothes: people recognize AI models quickly, and that breaks brand trust fast. So if you’re using it for marketing, make sure it’s always going through a human review process. And, ideally, not just the person who created it, because if you’re rushing to get something out, you’ll tell yourself it’s probably good enough. Have a second person check it, especially for a photo, so you’re not putting out something that makes people uncomfortable or distrustful of your brand.
Case Study: AI-Written vs. Human-Assisted Blogs
G.R.: I hear all the time, “I don’t need your services: we’ll just write our blogs through AI,” or, “I don’t need your social media services because AI will create it.”
Cory, you did a study on blog articles written solely by AI versus ones touched by a human. Can you talk about the results?
C.W.: It was very interesting. We did it on our own blog. I did a month of blogs written entirely by AI, and then another month of AI-assisted blogs, where I’d take the text it gave me and rewrite it in the style I actually write in. The number came out to roughly 3.7 times more views on the blogs written with human assistance compared to the ones written purely by AI.
So you can immediately see the cost of relying on AI to write content exclusively. Yes, it’s very good at scale and very fast, but you’re not going to get as much value, and over time relying too much on AI content is going to push a business further down in search results. Which we tend to forget when we talk about AI. Google is still setting records every month for the number of searches, so it’s not like Google is dying. People are still using it a lot to get information. SEO is not dead.
AI and SEO
G.R.: Great segue, SEO is a huge need for small businesses, and AI can both help and hurt it. Can we talk about SEO in general? I see people write a blog about, say, keeping foxes out of your backyard in winter, then copy and paste it straight onto their website. Why is that a bad thing? How can AI hurt SEO, and how can it help: keyword expansion, blog outlines, that kind of thing?
C.W.: It’s tough, because you can get a lot of content written through AI, but it’s not going to have the human creativity your brand needs to project in order to connect with customers. It’s going to sound pretty soulless and corporate, not in your brand voice. And, given how much of marketing today comes down to branding, that’s a recipe for disaster.
C.N.: The other point on SEO is duplicate content. If you’re putting in a prompt very similar to what someone else is using, and you’re not great at prompt engineering or training the tool, you’ll end up with duplicate content. Google will ding you for that, and it can also recognize when something is entirely AI-written, which it’ll also penalize.
So while it’s valuable for outlines and structure, it’s not useful for a complete write, Google will catch it every time. Cory proved that with his case study, because people don’t like reading things that are purely AI. You can spot it now, with twenty em dashes, a bunch of ordered lists, the same introductions every time.
So you can and should use it for certain things, like outlines and making sure you’ve got the right information, but you have to be careful about using it directly for SEO content, because it’s more likely to hurt you than help.
Google’s Stance on AI-Generated Content
G.R.: So what is Google’s actual stance on AI-generated content, and what does that mean in practice for small businesses?
C.N.: Google doesn’t want you to use purely AI-generated content. Cory, can you elaborate?
C.W.: This is a question we get a lot, and Google has actually addressed it. Google isn’t specifically against AI-generated content, which sounds surprising, but it’s more nuanced than that.
If you put a query into an AI tool and ask it for content, it’s going to give you content. And any other person or company that puts in a very similar prompt is going to get very similar content back. What does Google prioritize more than anything? E-E-A-T: expertise, experience, authority, trustworthiness. If you have twenty plumbers with nearly identical content about pipes freezing in winter, what does that tell Google? That there’s nothing helpful for the user in any of those blogs, because Google already indexed the main source the AI pulled the information from.
So, indirectly, AI-generated content on its own isn’t helpful. But if you do human assistance and add insights, regional information, and your brand voice, you’re adding real value to what AI produces, and giving it a better chance of indexing, which is what our case study proved.
But if you just take AI’s raw output and put it on your website, it’s not going to do anything. It might even hurt you with duplicate-content penalties, but at best it just sits there.
C.N.: Yep, he explained that much better than I was trying to! Google doesn’t want purely AI content because it doesn’t give people any actual value, and that’s the main thing: it wants you to provide value.
Practical Examples: Social Media and Ad Copy
G.R.: I know a lot of people came here for practical examples of how they can use AI. Can we walk through some real examples of how it’s successfully used in marketing?
C.N.: One tool I like, depending on what you’re using: a lot of our clients are probably familiar with HeyOrca, which we use for social media clients. It has an AI tool built in, and it also lets us schedule posts, which is itself kind of an AI piece people don’t always think about.
One practical use: if you come up with ideas using ChatGPT, you can pop those into a HeyOrca post, it gives you an AI prompt, you say, “I’d like to do a post about X, Y, Z, here’s the information,” and you can upload your brand voice there too. It takes the whole thing and spits out a caption with a few different options and emoji suggestions, and you go in and edit and tweak it. So it lets you write the post without having to sit and think through every single word… you give it the idea, it writes a draft, and you tweak it.
Another place it’s useful: going into ChatGPT directly. For example, I gave it this prompt: “Provide five social media post ideas for a digital marketing agency based in Baltimore. The agency provides full-service digital marketing with a focus on social media and SEO. The tone is a bit irreverent and playful while also being fun. This agency also prides itself on customer service and providing quality information to clients. Do not add any posts referring to X, as the agency doesn’t promote this platform.”
So I told it what I do, where I do it, how I do it, and specifically what I did and didn’t want. It’s a pretty short prompt as prompts go, so it’s not going to give me everything, but it will give me five solid ideas to work with — like “SEO MythBusters: Baltimore Edition,” a carousel busting common SEO myths local businesses hear, which is something we’d probably actually do.
So it’s a great idea generator and a great place to start. Then you take an idea you like and workshop it further by training the tool: “I like this idea, help me write a post based on it.” It’ll help you write the post, you edit it, and pair it with a graphic. That’s one way we use it to save time.
Cory, do you have some examples too? You’re a bit more on the data side.
C.W.: My biggest win with AI has been generating ad assets: giving me ideas for PPC. It’s done a great job helping me scale the back-end work I do on PPC by giving me a ton of headline, description, and caption ideas that I can review for compliance and brand-voice issues, local signals, and so on, then test within the ads platform.
It’s also, coincidentally, helped me write the first case study for our agency. I’d never written one before, so I asked ChatGPT for a rough outline of how to format it, and it gave me a good framework I used while writing the actual content myself.
Those are solid use cases for any small or medium-sized business.
A Quick Note on Jargon
G.R.: We’re using a lot of initials! How about we explain PPC and SMB?
C.W.: Pay-per-click, and SMB is small and medium-sized businesses.
G.R.: We get lost in jargon, and I never want to do that. I spoke at an event once, and someone complimented me on not using jargon, so we’ve got to stay clear on that.
C.N.: I started a job a few years before working here where they used initials for literally everything, and I remember sitting in my first meeting thinking, “I don’t think I learned anything today”, it was just initials.
G.R.: I know we can get lost in it sometimes because we live in it every day, so if you ever need us to slow down or clarify something, drop it in the chat and we’re happy to do that. No dumb questions, I promise.
Favorite AI Platforms
G.R.: Let’s talk about the platforms, since there are a million of them out there. I found one yesterday I might start calling my new friend, because she talks to me and I talk to her.
What platforms are you two using regularly, and why? Cory, you can go first.
C.W.: ChatGPT. It’s the most flexible platform, I think. A lot of good work has already been done by people much smarter than me building custom GPTs, and it’s just a very versatile system. If you’re deciding which system to try, just pick one and stick with it, because a lot of the tools operate pretty similarly, and you can transfer a lot of what you learn from, say, ChatGPT to Gemini. But overall, I find ChatGPT the most useful by far.
C.N.: I agree for marketing, I use ChatGPT a lot too. I like Gemini a little better sometimes for creative work. And a bonus tip, just for work in general: if you spend a lot of time in meetings, Read AI has been amazing. There’s also Otter and others out there, but I use it a lot to take notes and help draft follow-up emails, saving time. I still edit all my follow-up emails and double-check my notes against the actual conversation, but it definitely saves time and helps me remember things I might have missed.
So, as a bonus tip, Read AI is one of my absolute favorite AI tools right now.
What Businesses Should Stop Doing With AI
G.R.: I want to leave time for questions, so please start dropping them into the chat or the Q&A. While we wait, what are some things businesses should stop doing with AI right now?
C.N.: Letting it create graphics for you and just using them as-is. I hate using AI for graphics and images. I might use Canva to generate an AI image that I’m then going to manipulate and edit, but I’d never just go into ChatGPT and say, “give me a graphic.” I actually tried this recently for a client, just to see what would happen. I needed a timeline in graphic form, and it spit one out. I didn’t keep it or use it, but I sent it to Cory afterward, and it was one of the funniest things I’ve read, because every word was spelled incorrectly and in the wrong spots, even though the timeline layout itself was right.
It’s just never going to do a good job on graphics. It might give you a starting place if you know how to use something like Canva or Photoshop to build your own, but don’t rely on it purely for graphics and photos, because it does a poor job.
G.R.: I’d add: please don’t use AI graphics or images on your website. We have one client whose images are so clearly AI-generated. There’s a graphic of a fish on a plate, and it’s a color no fish in nature has ever been. Anyone looking at it thinks, “that’s definitely an AI image.” You’re going to get caught. And if I visit a website with AI-generated product images, it tells me the client doesn’t actually have a real product to showcase.
C.N.: I’d caveat that this mostly matters if you’re B2C or a B2B product provider. We have AI photos on our own site, but they’re clearly dogs. So there are times it works, but you have to be specific about it. Don’t try to AI a real product if you have one, because people will know it’s not real.
G.R.: Correct, and that was an e-commerce site selling food-related products, and the food wasn’t real food. It was fake food.
C.N.: Great explanation. Cory, what should people stop doing right now?
C.W.: Stop asking AI for content and then just publishing it without editing. It sounds bad to say, but it all reads the same. Same cadence, same narrative. I’m so tired of reading AI articles. I’m in the space, and even I think, “how can anybody read this and feel their five minutes were well spent?” I want that time back!
Practical AI Uses Beyond Marketing
G.R.: What are one or two smart, practical uses outside of marketing? How else are you using AI day-to-day, in your job or personal life, that people could try today?
C.N.: I mentioned Read AI, which I use a lot for work. It helps me write follow-up emails. I also use it to help create task lists and figure out next steps.
For example, I had a two-hour meeting with a client recently that covered a bunch of changes they wanted made. I took my own notes, but I also had the transcript from Read AI, so I copied it into ChatGPT and asked, “What are the action items from this meeting?” It listed them all out, so between my own notes and anything I might have missed on a two-hour call, I had a complete picture of what the client asked for.
On a personal level, a family member of mine had a serious health scare recently, and we were trying to figure out what questions to ask the doctor, what our next steps were… legal questions, all of it. I’d never dealt with anything like that before, so I had a lot of questions. I asked ChatGPT what questions I should ask the doctor based on the diagnosis, and what we should look at next, and it gave me a bunch of genuinely helpful things I wouldn’t have thought of otherwise. It was faster than doing a Google search and reading through five or ten different websites. It’s useful for aggregating information quickly when it’s something you don’t know much about but know there’s a lot of information out there on.
C.W.: I’ve got two. First, product recommendations. I’ve used ChatGPT to help me shop, mainly because you can give it a lot of qualifying information that you just can’t give Google. I gave an example in a past webinar: I was shopping for a backpack, and I gave it specific requirements: a certain material, needed to fit a 16-inch laptop, good for travel, under $150… and got good recommendations, because ChatGPT and other AI tools pull from customer reviews across sources when making them.
Second, I’ve used it for any search where Google didn’t feel like it was going to cut it. For example… you all know I’m moving to Japan in January. I had a question about bringing enough prescription medication for my stay, and Google was giving me a lot of old websites and outdated information. So I asked ChatGPT directly what to do about my prescriptions for the trip, and it actually gave me the immigration authority’s website and all the information I needed, plus a step-by-step timeline for how to handle it. It was massively helpful.
Using AI to Research Your Own Business
G.R.: One more marketing tip before we wrap up, which Cory and I have talked about on previous webinars: one of the best things you can do is ask ChatGPT about your own business. Go in and ask, “What are people saying about my business? What similar businesses are out there, and what are people saying about them? What are they doing that I’m not?” Look for the gaps. Just talk to it. Ask if you’re showing up for certain things. That’s kind of the beauty of ChatGPT, Gemini, or any of these engines: you can ask them questions about your business, and they’ll give you a sense of where you stand across Google reviews, social media, and other platforms. So it’s a good idea to just go in and ask what it knows about your business.
C.W.: Piggybacking off that: one thing you can do is put yourself in the shoes of your ideal customer and ask the question that person would ask ChatGPT. Going back to the plumbing example: ask, “What are the best plumbers for burst pipes in Baltimore?” It’ll give you a list. Then go back and ask, “Why didn’t you recommend my company?” – put in your company name, and it’ll tell you, maybe you don’t have enough reviews on Google, or it’s not clear from your website content what you offer, or there’s not enough social proof. It’ll basically give you a roadmap for what to improve.
C.N.: One exercise we did while building our new website: most people say they hate the content on their website. Too much, too little, doesn’t say what they do, hasn’t been updated. We put our website and three competitors into ChatGPT and asked it to compare and contrast the content in terms of messaging, reliability, and a few other factors. It came back with, “you talk about this, but they talk about that.”
It was interesting: our website talked entirely about customizable solutions, while the competitor sites focused on how long they’d been around. Which is fine, but it doesn’t really give anyone a reason to buy from them beyond longevity. It doesn’t mean they’re taking the time to give people the solutions they actually need. So doing a compare-and-contrast between you and your competitors can show you where you’re missing messaging, or where they are, that you can capitalize on.
Closing Thoughts
G.R.: We’re just about out of time, and I know we haven’t hit everything people in the room wanted to cover. So I’d encourage you to email me if there are things you’d like us to talk about, whether AI or anything else. We want these to be valuable to you, not just us talking about what we find interesting. If there’s appetite for an “AI 2.0,” we’d be thrilled to customize a future webinar around that. We are recording this, and it’ll be on YouTube for later viewing, whether you were answering emails while listening, had to leave early, or want to share it with someone else. And if you have specific questions about your business, your marketing, or your AI use, we’d love to have that conversation with you.
Thank you all so much for attending: I really appreciate everyone who came out to watch and listen today. Happy to answer more questions, and if you’d like to see a part two, let us know. We’ll probably schedule one for January; not sure about December with the holidays. We love talking about AI, so we’re always happy to do it.
Cory, any final words?
C.W.: Just be careful with it. It’s very good at accelerating what you do, but don’t trust it blindly, and verify everything.
G.R.: Agreed. Thank you all. If we don’t see you before then, I wish you happy, healthy, wonderful holidays, and we’ll see you maybe in December, but definitely back in January. Thanks, everyone, take care.
