Mastering AI Prompting: Advanced Techniques for Success

My first real aha moment with AI in marketing was when I discovered and personalized a prompt sequence that enabled AI to help me prepare for podcast interviews, trimming my time down from two hours to just twenty minutes.

Creating effective AI prompts is an art and science that can drastically elevate the capabilities of AI models. While many are familiar with the basics, mastering advanced techniques like chain prompting and building custom GPTs can unlock new levels of functionality and precision.

If you’re looking to deepen your understanding of AI prompting and achieve greater success with your AI projects, this episode is for you.

Today, we’re joined by Ashley Gross, an expert in teaching people how to build businesses using AI. Ashley will share her advanced techniques for AI prompting, including best practices, tips for chain prompting, and insights into building custom GPTs. With her extensive background and certifications in AI and data science, Ashley is well-equipped to guide us through these advanced strategies.

The AI Hat Podcast host Mike Allton asked Ashley Gross about:

✨ Mastering Chain Prompting: Learn what chain prompting is and how it can improve AI responses.

✨ Custom GPTs: Discover the process and benefits of building custom GPTs for specific business needs.

✨ Advanced Prompting Best Practices: Gain insights into best practices and advanced techniques for effective AI prompting.

Learn more about Ashley Gross

Resources & Brands mentioned in this episode

Full Transcript

(lightly edited)

Mastering AI Prompting Advanced Techniques for Success with Ashley Gross

[00:00:00] Ashley Gross: I actually don’t see prompting going away as fast as everyone else keeps saying it because there’s so many people that aren’t great at communicating and they’re not, even though a lot of people know about AI and they’ve tried AI, they’re not getting benefits or real value from AI. Like we haven’t hit that milestone yet collectively that I think it would be safe for companies to think about like taking away prompt engineering and having some kind of an interface already built in.

Because then again, you’re assuming that people know what to do and how to even use those directions. So I don’t think the prompt engineering is going anywhere yet.

[00:00:35] Mike Allton: Welcome to AI and marketing unpacked, where we simplify AI for impactful marketing. I’m your host, Mike Alden here to guide you through the world of artificial intelligence and its transformative impact on marketing strategies. Each episode, we’ll break down AI concepts into manageable insights and explore practical applications that can supercharge your marketing efforts.

Whether you’re an experienced marketer just starting to explore the potential of AI, this podcast will equip you with the knowledge and tools you need to succeed. So tune in and let’s unlock the power of AI together.

Greetings program. Welcome back to AI in Marketing: unpacked where I selfishly use this time to pick the brains of experts at keeping up with and integrating or layering artificial intelligence into social media, content, advertising, search, and other areas of digital marketing, and you get to learn to subscribe to be shown how to prepare yourself and your brand for this AI revolution and come out ahead.

My first real aha moment with AI marketing was when I discovered and personalized a prompt sequence that enabled AI to help me prepare for podcast interviews, trimming my time down from two hours to just 20 minutes. Creating effective AI prompts is an art and a science that can drastically elevate the capabilities of AI models.

While many are familiar with the basics, mastering advanced techniques like AI, Chain prompting and building custom GPTs can unlock new levels of functionality and precision. If you’re looking to deepen your understanding of AI prompting and achieve greater success with your AI products, this episode is for you.

Today, we’re joined by Ashley Gross, an expert in teaching people how to build businesses using AI. Ashley is going to share her advanced techniques for AI prompting, including Best practices, tips for chain prompting and insights into building custom GPTs with our extensive background and certifications in AI and data science.

Ashley is well equipped to guide us through these advanced strategies. Hey, Ashley, welcome to the show.

[00:02:29] Ashley Gross: Hi, thanks for having me.

[00:02:31] Mike Allton: So glad to have you here. Let’s just dive right in to start. Could you share a bit about your journey into AI and why you focused on AI prompting techniques?

[00:02:41] Ashley Gross: Sure. I’m happy to. Prior to 2020, I had no technical background, no technical experience. I simply wanted to use AI to help me cut down on my workload because I was expecting my son and I knew I had to get really, really scrappy with my time and prioritize it. So I started using Jasper to help me with my email tonality and just unifying my brand voice.

And then once I discovered. How fast I was able to automate that. I was like, okay, I need to start really cutting down my workload and trying to figure out everything that I don’t like to do one and then two, what takes up the most time and how can I automate those tasks to give me more of my time back to spend time with my son.

So I. Obviously figured out how to cut down a 40 hour work week into, you know, 15 hours and then 10 hours. And I was very vocal about it. I started telling everybody at my enterprise what I was doing and I was like, try this, do this. And I was so excited. It was like, I found the work balance. And soon before.

Back to our CM, she, you know, asked me if I wanted to lead an AI task force and roll out generative AI to an enterprise marketing team of over a hundred marketers. So as soon as I realized that I was going to have to make sure that every single person had the best possible experience with AI, I realized that the number one thing that I need to focus on was prompt engineering.

And that was because I couldn’t control What they were putting into the AI or how they were using AI, but I could control their learning experience and how they first learned to use AI. And I thought, okay, the basics of that is prompt engineering. If they can get really, really good at communicating what they want and what the outcome is and how to get from point A to point B.

Then if there’s any mistakes happening with AI, it’s probably not prompt engineering and I can fix those issues. But if it’s an issue of prompt engineering and they’re not learning how to do this from day one, I’m never going to be able to roll this out effectively. So that’s kind of how I fell into prompt engineering and putting an emphasis on it.

[00:04:39] Mike Allton: Love it. And I love that you stress that it’s about communication, which I’m sure we’re going to talk about more. But for those who haven’t really done this very much, maybe they’re brand new and they’re just starting to think, you know what, I do need to start integrating AI into my work or my own professional life.

Can you just briefly explain what AI prompting is and why it’s so important?

[00:04:58] Ashley Gross: Yeah, it’s just giving the AI instructions to help it get better useful responses back. So think of it as having a good conversation where if you ask really great questions, then you get really great answers. But if you don’t ask really good questions, then you get crappy outputs.

You want to make sure that you’re guiding the AI to deliver The most relevant and valuable insights. And it’s really just about getting the most out of that AI tool.

[00:05:23] Mike Allton: That makes a ton of sense. Hopefully that makes sense to all of you that are listening. So what are some fundamental best practices for creating some really, really good, effective AI prompts?

[00:05:34] Ashley Gross: Yes. These sound really simple, but I don’t think that they can be overstated enough. So be clear, know what your prompt is and make that prompt extremely clear in order to avoid confusion. You need to provide context. So give AI enough background information to understand what you’re asking it to do. You need to be concise, which I know kind of sounds counterintuitive, but if you want to, and I know we’re going to get into this, but if you want to give multiple prompts, try not to smash it all into one prompt, because if it’s one and it’s super long, then it’s not going to do as good of a job.

It’s going to be overloaded. The other two basic foundation pieces I have are to test and refine continuously and then use feedback. It’s a lot easier now that chat GPT has that feature automatically. So you don’t necessarily have to prompt it, but it’s always good to just make sure that you’re giving it your feedback and saying like, okay, do you understand what my feedback is and how to move forward?

And then they acknowledge it. And then you can move forward.

[00:06:31] Mike Allton: That’s great advice. Just understanding that you can and should revise what you’re asking the AI, revise the output that you get. And I often tell people, look, you can’t break it. So don’t worry. You’re not going to, because. The way I relate to it is a little bit personal.

I have a value of time. In other words, my top three values are love, friendship, and time. With people waste my time, that irks me. That’s a pet peeve of mine. And I recognize that and I defy that. But if somebody asked me the same question six times in six different ways, I would start to lose patience with them and they would probably hear that my voice and see it in my face, but the A.

I. Is never gonna do that. You can ask it the same question 10 times and you could ask it to give you 10 more answers with each iteration of that question. And it’s happily saying absolutely. Ashley, here’s 10 more titles for you to think about. And it’s so happy and cheerful. So that’s fantastic. But You kind of talked about a topic.

You said, don’t smash all the prompts together. Let’s talk about the opposite of that, right? That would be the opposite of that would be chain prompting. What does that look like exactly? And how does it improve the AI’s responses?

[00:07:44] Ashley Gross: Yeah. So chain prompting is just breaking down one big question into a series of smaller questions.

And it helps the AI handle complex questions or tasks that you’re giving it. It helps it be more accurate and it gives more detailed answers because you’re taking the time to make sure that instead of giving it one big task, you’re giving it a little piece and making sure it understands. Give it a little bit more, give it a little bit more, you know, breadcrumbs, because if it doesn’t understand one aspect of it, then it’s not going to give you a great output.

Then you just wasted all that time. Adding like six questions into one prompt, one, if you did it individually, it would have just understood it better because that’s how it’s trained as an algorithm. It’s trained to predict the next word. So if you’re smashing all that together, it’s like inundated with information and it’s like not computing.

It’s like what do you want me to do next?

[00:08:31] Mike Allton: And my friend Dustin who’s a founder in the AI space talks about how the default Mode or persona of AI is to be helpful. And we see that when we say things like, Hey, I’d like some ideas for a blog post title, and the AI gives you 10 blog post title ideas, and then goes on to write an outline for you and draft the introduction for you.

And you’re like, Whoa, wait, I’m not even happy with these titles yet. You’ve moved too fast. And so I think correct me if I’m wrong, when you’re doing these chain prompts, you need to specifically say something like. And then stop, right? You need to tell the AI to read whatever it is you’re telling it, provide some information or context or answer a question, and then stop for more feedback, wouldn’t you say?

[00:09:15] Ashley Gross: Yeah, I would say I always tell people like I have five rules because I just I had to train people how to do this. So I took a process and I simplified it again, I simplified it down until it was bare bones. So I always tell people to input five rules into an LLM. Before you start to use it. And those rules help it mitigate hallucinations and it help it to actually not generate outputs unless you fully fill in the input. Oh, I put a beginning. So you just have to think about it once, implement those rules, and then make sure you’re iterating and checking it for accuracy later, that helps a lot more than if you were just to go into this and have to, you know, manually recall the memory.

[00:09:53] Mike Allton: Can you share those five rules?

[00:09:55] Ashley Gross: Yeah.

I’m happy to

[00:09:58] Mike Allton: teasing them. There’s five rules, but you can’t know what they are.

[00:10:02] Ashley Gross: No, not at all. So I would say the five rules are you have to tell ChatG PT, like I’m going to give you rules. You have to follow these rules. You’re not allowed to generate an output unless the input is fully filled in.

And if I don’t fill it in, then you need to remind me. So it’s just making sure that it doesn’t generate an output with a half thought input. Tell the GPT to validate the resources and accuracy of the data to prevent baseless conclusions. So if you tell it something, essentially what this rule does is if I say the capital of I don’t know,

is Berlin or whatever, if I make up a blatant lie. The rule then tells ChachiPT to come back to me and say, actually, that’s not true, da da da da da da da, because I’m not a perfect human, now I know the capital of France, but I’m not, I’m going to mess up sometimes, so I need it to hold me accountable as well.

And then the other rules are verify that you understand the context. So after every input, ask a question. I like to do it based off of time. So I like to say after 30 minutes or after an hour, you need to verify the context of what I’m asking you to do. Just because if I’m on there for a long period of time, I want it to remind me of like, Hey, are we still on this task?

Is this still the context? And then now it automatically does this, but telling chat GPT to store your data in the knowledge base so you don’t have to continuously remind it not to say delve in queries. And then there’s one more rule. Ask questions. So I, I teach my to understand the request. So if I say, tell me about Washington, then my AI sits back.

Do you mean Washington, D. C.? Do you mean Washington, the state? Like it asks questions before it continues to answer it. I just found that putting those rules into place early on makes it a lot easier. Easier to learn it. I mean, I feel like if you spend a ton of time getting onboarded to a tool, did you actually use it for productivity or was just one more thing that you had to learn?

[00:11:54] Mike Allton: Those are fantastic rules. I know probably within 12 to 18 months, a lot of those won’t be necessary. Like we already know with chat GPT’s capacity it’s already understanding context a lot longer and a lot more and a lot deeper. And so as we go, it’ll stop using words like delve, it’ll stop trying to start every paragraph with in the vast landscape of whatever it’ll move on from those.

It’ll learn, it’ll get better. But yeah, having to, you know, having those rules in place to train the AI right up front. Fantastic advice. Can you share maybe a specific example of how you’ve used chain prompting in one of your projects and how that worked out for you?

[00:12:33] Ashley Gross: Yeah. So I built a sales GPT. And the GPT had three templates inside of it.

So one was outreach sequences, one was competitive intelligence, and then one was market research. Because when we took a poll of our whole entire sales organization, they said that those are the three activities that were hogging the most amount of their time. And so my chain prompting looked a lot like I would write down what the sales reps process was already.

So like in the mind of a sales rep, they would go do. You know, market research, right? And they would figure out how our product differs from whatever our prospects product is. And then they would go in and do some competitive intelligence, like on the actual account. And then they would start to create their outreach sequences for emails and SMS, you know, and you get my point.

So when I was prompting it, I was saying, Hey, you’re an SDR. Do you understand? Yes. Great. Here’s our product. Do you understand? Yes. Great. Then it was, you know, when I’m looking at targeting a mid market account, I need eight to 10 reps in there and they need to be the personas of X, Y, Z. Do you understand?

Yes. So it was a lot of just taking them through what that process actually looked like in real time so that they understood what the workflow was like, because that’s all you’re doing is you’re teaching it. your workflow so that it gives you a solution faster than if you were to manually do it, but you’re not changing your workflow.

And it makes it a lot easier that if the AI gives you an answer that doesn’t necessarily make sense, if you gave a chain prompt and you took her through the scenario I just took you through, where it’s like, here’s what you do, here’s what you care about, here’s who you are, here’s who we’re targeting.

Then if she gets caught up at a certain part, then you know a certain part and you don’t have to start from zero. If you were to just put all that in one. It would probably try to like put all that into one template and it would just be a terrible template.

[00:14:22] Mike Allton: That makes a lot of sense. We’ve called that pre work before.

And some of our past episodes with folks like Andy Christodina and Chris Penn and so on. So you can listen to those episodes a little bit more, but we stress how important it is to spend that time up front so that you ensure the outputs you get. And not only correct, but they’re going to be fast and more efficient.

And then you can replicate them and scale them over and over again. Folks, we’re talking with Ashley Gross about how to build the best prompts so that when you’re working with AI, you get the best possible results. And I’ve got a bunch more questions, but before I get to those, let me share with you the tool that I’m using every single day to help me with my AI.

This episode of AI and Marketing Unpacked is brought to you by Magai, your gateway to making generative AI, incredibly simple. Wondering how to seamlessly integrate AI into your marketing strategy without getting bogged down by complexities. That’s exactly where Magai shines. It provides user friendly AI solutions that empower marketers just like you to innovate and elevate your campaigns without needing a degree in science.

Imagine having the power to generate creative content, insightful marketing data analysis, or even personalized customer communications all at the touch of a button. Magai isn’t just about providing tools, it’s about Transforming your approach to marketing with AI that’s tailor made to be straightforward and effective.

So whether you’re looking to boost your content creation process or want deeper insights into your marketing performance, Magai makes it all possible with a few clicks, no fuss, no hassle, just results. Ready to simplify your AI journey? Visit Magai today to learn how their solutions can revolutionize the way you engage with your audience.

with your audience. Don’t just market, market smarter with Magai. Tap the link in the show notes. So Ashley, building custom GPTs, that kind of sounds complex. That sounds like something I might need a computer programming degree to actually wrap my head around, let alone actually do. Can you explain the process?

What is a custom GPT? What are some of the benefits and how do we set them up?

[00:16:25] Ashley Gross: Yes. I’m to miss miss. Myth myth bust. That’s the word I was trying to say. I would love to myth bust that real quick. Because it’s really not difficult. And if you think about the amount of time that you’re putting in to either build a GPT or learn a GPT, if you know what you want and how to get from point A to point B, it takes a lot less time to actually build your own GPT for that specific tool than to use an all in one platform that maybe has the features that can solve your problem, or maybe it’s just going to be more.

of you researching more aspects of that tool and hogging up more of your time. So that’s my myth busting TED talk, but yeah, so the process is. You just define your goals like you would any other task that you’re assigned. You gather the data that you need to make sure that you can actually accomplish that using AI, because that, you know, a lot of times is where people get hung up.

You cannot solve everything with AI, no matter who tells you that. You design the prompts, and I recommend that people design prompt shortcuts as well, because if you have a couple of people that are beginners, but a couple of people that are, you know, more intermediate expert level, if you level the playing field by adding prompt shortcuts, and what I mean by that is the ability to say like prompt forward slash one means X, Y, Z, so that people don’t have to manually copy and paste from a word doc and do a GPT, it makes it easier.

So design prompts and prompt shortcuts. Train and fine tune and then test. I cannot iterate that enough though. Like test, test, test, get real world scenarios, get feedback, iterate, and then roll it out to everyone else. But as far as the benefits, I mean, it’s huge. You’re going to get highly tailored solutions in it.

And I mean, obviously, you know, improve efficiency. It’s going to tackle very specific business challenges more effectively because it’s not using a generic model. I mean, I also tell people that want to build a custom GPT. Just to like hang out in the GPT store and find a custom GPT that they want to test just to see like the difference between the generic and the specialized I just starting coloring book hero, it’s a GPT made by chat GPT and it just allows you to create any coloring page you want off of a prompt that you create.

So my toddler loves dinosaurs, so I’m able to just create a coloring page for him in chat GPT in under five seconds. So like being able to see what one specialized GPT does and then trying to do that in just chat GPT four on your own and seeing the difference in time it takes you and how the amount of prompting.

Okay. Bye. That is a good bench line for people that want to start.

[00:18:47] Mike Allton: Okay. First of all, the coloring book hero tip that is huge. I literally wrote that down. I’ve got a nine year old daughter, two girls, but my youngest is nine and she loves coloring and so every couple of days she’ll walk into my office and say, daddy, can you print me out some X dragons, mermaids, whatever.

And then, you know, then I spent five minutes, 10 minutes on Google. Searching for coloring pages and trying to find ones that are the right sizes and that she wants. And that’s a whole thing. So that’s first of all, thank you for that, Ashley. And I’m glad you mentioned the GPT store. It hasn’t, it’s not something we really talked about on the show yet, but folks should know if you’re using open a eyes chat GPT, there’s a whole store where you can get, you know, free paid custom GPTs that businesses had developed.

A lot of them have their own websites and their own businesses around. If you’re using Magi, like we talked about in that commercial break, Magi refers to these as personas. They operate exactly the same way. You’re putting basically a chain of prompts inside there and telling the AI step by step what to do.

I mentioned at the outset, I have one right here. For preparing for these podcasts. I tell the podcast, Hey, I’m talking to actually gross. Here’s her LinkedIn. Here’s what we’re going to talk about. And it throws me some ideas for topics and then some topic some actual episode title ideas and questions and everything.

Then after my podcasts come out, I have a whole separate. Custom GPT or persona that I built that helps me create my promotional campaign for the full week after the episode Drops, I’ve told it, you know who I am what the episodes are about the topics the audience the channels and the cadence I want to use and helps me develop all the social posts a whole LinkedIn newsletter everything for a whole week It pulls out quotes and everything.

It’s fantastic. So that’s terrific What, Ashley, would you say are maybe some more advanced techniques for getting these kinds of custom GPTs in our hands, meeting specific use cases?

[00:20:41] Ashley Gross: So definitely chain prompting, like we were talking about, that’s really important. I would also say it’s less of an advice and more kind of just.

general guidance, whatever you put into the knowledge base takes precedence over the search engine. And what I mean by that is if you were to go in there and say, what does the prompt community do? But then in your knowledge base, it just had information about your business. Let’s just say the knowledge base would take precedent and it would keep trying to pull from that knowledge base instead of using the search functionality.

We use the knowledge based functionality because we only wanted it to know certain things and we didn’t really need it to scan the web because we had, you know, quarterly revenue, quarterly annual review all these different documents that we could feed it so we knew it was trained on the accurate data.

If you do want something that is keeping up with statistics, like don’t put anything in the knowledge base and just use the search functionality and spend your time working on the specific rules for the search engine. I would also say like, if you don’t want to do any of that because it just sounds like a nightmare, you can attach things.

Connect your Google drive. So that’s another great way to create the GPT is you can copy and paste any URL and say like, here’s everything you know about this URL. This is what I need you to do. Or you can just connect your Google drive, assuming that there’s nothing personal on that specific drive.

And then that’s its whole knowledge base, which is also a great way to go. If you’re going to use an LLM, like Jasper, for instance, because it’s not trained off of that. So you get your Google drive and then you have. You know, all of your marketing content and all of your marketing templates and you can just like consolidate all of your information.

So when you’re asking the LLM for something specifically, I could say, create me a blog using, you know, company XYZ is framework and fill in the template so that I can present it to the CMO for approval. Right. And it could do all of that stuff because it had all of that information already connected. Sorry, I went nerdy really quickly.

Let me get that

[00:22:35] Mike Allton: on this show. We embrace nerdiness.

[00:22:38] Ashley Gross: And then I would just say like play around with integrating the GPT once it’s pretty accurate on its own. Like you can plug and play APIs, which make it so much better. You can integrate it with other tools to actually like automate a workflow for you so that the answer doesn’t just like live and die in chat GPT.

So like test yourself and there’s not a lot of really truly interesting use cases out there because people are experimenting. That doesn’t mean that we should keep going for like the low hanging fruit use cases for AI because it’s like so much more than that and it can do way more. So. Whatever process you want automated, I’m sure you can do it.

[00:23:12] Mike Allton: Yeah. And I’m glad you mentioned Jasper friends of the show, Jessica Hreha from Jasper will be on the show in a little bit. Folks might not realize they’ve been around for a long time. They used to be conversion. ai. So they, they’ve been very forward thinking in the AI space for a while. So we’ll do a whole show about what you can do and what you can’t do with, with Jasper.

So thanks for sharing that. But I think the whole point of doing custom GPTs is so that we Have them help us have the A. I help us with tasks that are going to be replicated and scaled over and over and over again. So how do we ensure that it’s accurate? How do we ensure that it’s consistent when we’re using some of these advanced A.

I propping techniques?

[00:23:49] Ashley Gross: Yes. So I would say obviously the five rules that I went over and put those immediately before you start playing around with AI, and then I would also just say, like, this is impossible without having a human the real you as a user to make sure that you’re staying up to date, you know, if you know that a new feature update is coming out, it’s your job to know like how that’s going to affect that tool because you’re using that tool.

So if you want your information to. be accurate and up to date. You need to also make sure that you’re training yourself every time these updates come out, which is why I say, like, pick and choose which a I told you on board yourself to because you can’t pick all of them. That would be impossible. And you’d stretch yourself then.

But that’s what my advice would be.

[00:24:29] Mike Allton: I want to go back to something you said almost at the outset of the show, because you talked about how AI council at your enterprise work. Can you share more about what that was and what it accomplished?

[00:24:41] Ashley Gross: Yeah, it was really just, I remember a slack message going out saying like, who’s excited about AI?

And it was our chief strategy officer that sent it out. And I was like, well, yeah, me. He’s like, okay, great. Who else? And like, we got a couple other people. And I think that the way so nice. We, like, this was pre chat GPT, so we had no idea of what was about to happen. It was really just, we don’t know if the reward’s going to pay off, but we’re going to play around with it and we want to get some, you know, ROI metrics.

So I think, yeah, I just, I think

just start. Just start somewhere and learn a little bit and get together with a collective group of people who all want to see impact driven and have individual perspectives on how that can get done. Right. And then align on one goal and then just like tackle it as a whole team and throw everything you have at it.

And. Like you’ll see a lot get accomplished. Like I said, we had over 25 AI use cases within the first three and a half months when we rolled AI out. But what was more impressive was we were given a pipeline target of 90 million and we achieved it. We overachieved it and we hit 115 million. So like we had quantitative and qualitative metrics.

You know, if you think about like what impacts you want to make, you can have multiple impacts on the way to one joint unified outcome. So I would just say like, get together with a group of people. With one common goal. Like what is the business’s problem this quarter? And it doesn’t matter where those people come from.

As long as the one thing that they have in common is they’re excited and they’re eager to learn about AI and like really give this a good shot.

[00:26:22] Mike Allton: I love that. And the reason I asked you that is because we were stressing how much. We need to keep people involved in these processes. And that goes to the very tactical level of building the prompts, but it also goes to the higher strategic level of how involved are people in the decisions of how we’re using AI in this entire organization, right?

You can, I talk to businesses all the time and. They’re often ignorant of how much their individual employees are or are not using a I because they’re not having these conversations. So we talked a minute ago. You suggested how you can plug in your Google drive as a link. And you very smartly said. Be careful.

You’re not having personal proprietary information in that drive link before you share it, that’s not something people necessarily know unless they’ve been told, right? So they could go to the GPT store. They could see, Oh, wow. I could get this custom GPT. That’ll shave my time at half, not realizing that they’re putting their Jeopardy by sharing proprietary information because I just don’t know.

So there’s, there’s my Ted talk and my PSA for all of you listening, start having these conversations today inside your organization. If you’re in charge, if you’re the CEO, you’re the chief strategy officer or CMO, brilliant. Take the leadership today to form an AI council. If you’re not, if you’re lower down in the wrong building, you can start having those conversations with your manager.

And start warming them up to the idea that, yes, we need to invest a little bit of time and manpower and having those kinds of conversations. Thank you for kind of queuing up that conversation, Ashley. You’re welcome. I just have a couple more questions. The one, we’ve already talked about Jasper. What other tools or platforms do you recommend for those that are really wanting to get into what we would call advanced AI prompting?

[00:28:04] Ashley Gross: Can I answer this in a weird way?

[00:28:07] Mike Allton: You absolutely can. And I, I hesitate before I say yes, but go ahead. Answer it in a weird way.

[00:28:12] Ashley Gross: No, no I, I feel like it’s a great answer. It’s just, instead of saying like what tool I recommend, I think that, I don’t think people should approach it that way. I think that people should look at things that they need to do every day, every week they hate doing.

And then go try and find AI solutions to solve that problem that they hate doing. And when they get to the point where they have like two or three vendors that they’re deciding behind or between, I would say like reach out to your network of people, see if anyone else has tested one of those two to three vendors themselves.

And if they have any feedback for you, try to buddy up if you can, and then go learn about that tool. Because like I said, There’s, there’s so many tools, it’s like impossible, even if it’s not an AI tool, to know everything that there is and know every feature and how to utilize everything.

[00:28:56] Mike Allton: I couldn’t agree more on one of the other shows that I do the Martech show. We brought in Scott Brinker, who is the chief ecosystem officer at HubSpot, and he’s also the editor at chiefmartech. com. He does an annual marketing technology landscape review. And there was a. The 23 or 28 percent increase in marketing technology apps from last year to this year, tens of thousands of apps driven obviously by AI.

And he’s not even counting all the custom GPDs. He said, well, if they’ve got a website, we’ll talk about it, but maybe in years down the road, we’ll just have to count them all. It’s insane. How much The software landscape is exploding. There’s just so much opportunity out there, which means to your point, there’s so much to look through.

There’s so much confusion, unsure, and there’s costs. That was one of the big points that we made in that conversation. And I’ll link to you in the show notes for you guys to go listen to this conversation with Scott. The whole point was be careful how much money you’re investing and how many different apps and programs and websites you’re signing up for, particularly if it’s just one simple use.

When to your point, instead of thinking about the tool first, think about, well, what are we trying to accomplish? What’s the pain? What is it that I’m doing today that I’d really rather not have to spend a lot of time on, if any at all. And how can AI help me with that? Love that approach. Now, my last question, Ashley, you’re just going to love because things are changing pretty quickly when it comes to AI.

And yet, despite that, I want you to put your Nostradamus hat on and tell me what’s going to happen in the future. What trends do you see, particularly with AI prompting? And this is actually an important question because some of the conversations I’m having with other folks who are in this revolution, they’re really paying attention.

We’re wondering how long are we even going to use? Prompting, right? Is prompting going to go away? Is it going to be replaced by future trends and technology? What do you think? Where are we going with all this?

[00:30:51] Ashley Gross: I actually don’t see prompting going away as fast as everyone else keeps saying it, because there’s so many people that aren’t great at communicating.

And they’re not, even though a lot of people know about AI and they’ve tried AI, they’re not getting benefits or real value from AI. Like we haven’t hit that milestone yet collectively that I think is It would be safe for companies to think about like taking away prompt engineering and having some kind of an interface already built in because then again you’re assuming that people know what to do and how to even use those directions.

So I don’t think that prompt engineering is going anywhere yet. I do see AI systems becoming more personalized individually and like highly tailored with the responses. I definitely think that’s going to get better. But I also see like the integration of AI with business tools that already existed.

I think a lot more companies that have been around for a while that we’re used to seeing as far as, you know, marketing tools, automation tools, sales tools. To add an AI functionality is because it’s still easier to be an established business and add in the AI piece already having that brand reputation than to start net new and have to gain the respect of everyone while being solely an AI product and AI products haven’t.

been publicly around for that long, if that makes sense. And then the other aspect is like, maybe this is like a wish list of mine, but I also think I’m seeing more of a trend towards it. So ethically, I use, I think it’s going to be a major focus and there’s going to be more attention on like the transparency, reducing bias and showing fair outcomes.

I think that We’re obviously going to make it easier for people to interact with AI without needing technical expertise, but they’re going to be, you know, adapting and continuous over time. So who really knows?

[00:32:37] Mike Allton: No, the point on bias is spot on. It’s something I even talked about in a recent newsletter article, and I’ll link to that in the show notes because that’s a really, really big deal.

So we’ll be talking about that a lot more in episodes to come. But actually, We’re out of time and this has been absolutely amazing. You’ve been such a treat to talk to. So so much value. So thank you for sharing all that for folks who want to learn more, where can they go to learn more and connect with you?

[00:33:01] Ashley Gross: You can go to get the prompt. com. That is where my prompt community is and that is where I live seven days a week.

[00:33:11] Mike Allton: Fantastic. We’ll have that and all of other, all of Ashley’s other information in the show notes. And that’s all the time we’ve got for today, friends. Don’t forget to find the AI in marketing podcast.

On apple and leave us a review. We’d love to know what you think until next time. Welcome to the grid. Thanks for joining us on AI in Marketing: unpacked. I hope today’s episode has inspired you and given you actionable insights to integrate AI into your marketing strategies. You enjoyed the show, please subscribe on your favorite podcast platform and consider leaving a review.

We’d love to hear your thoughts and answer any questions. Any questions you might have, don’t forget to join us next time as we continue to simplify AI and help you make a real impact in your marketing efforts. Until then, keep innovating and see just how far AI can take your marketing. Thank you for listening and have a fantastic day.

Related