GenAI as a CMO’s Strategic Ally

Drew: Hey, it’s Drew. Welcome to another episode of Renegade Marketers Unite. This show is brought to you by CMO Huddles, the only marketing community dedicated to B2B greatness, and that donates 1% of revenue to the Global Penguin Society. Wait, what? Yeah, it’s kind of weird, isn’t it? But let me explain. It turns out that B2B CMOs and penguins have a lot in common. Both are highly curious and remarkable problem solvers. Both prevail in harsh environments by working together with peers, and both are remarkably mediagenic. And just as a group of penguins is called a Huddle, our community of over 400 B2B marketing leaders huddle together to gain confidence, colleagues, and coverage. If you’re a B2B CMO who can share, care, and dare with the best of them, do yourself a favor and dive into CMO Huddles. We even have a free starter program. Now let’s get to the episode.

Narrator: Welcome to Renegade Marketers Unite, possibly the best weekly podcast for CMOs and everyone else looking for innovative ways to transform their brand, drive demand, and just plain cut through. Proving that B2B does not mean boring to business. Here’s your host and Chief Marketing Renegade, Drew Neisser.

Drew: Hello, Renegade Marketers! You’re about to listen to a Bonus Huddle, a specially curated Huddle that we run once a month with experts sharing their insights into the topics that are most important to our CMO community. We call them Huddlers. 

At this particular Huddle, we were joined by Liza Adams, Managing Partner at GrowthPath Partners. Liza will walk you through how she’s using generative AI to rethink the entire strategic process and generate segments and personas. It’s really fascinating. Let’s dive in and see how Gen AI is reshaping strategic planning. 

Hello Huddlers, I’m excited to welcome you to our seventh Bonus Huddle focused on some aspect of generative AI. Last month, Andy Crestodina walked us through the art of training your AI. If you missed that one, be sure to track it down or just put it in chat and we’ll be sure to send you the link. This month, our very special guest is Liza Adams, Managing Partner at GrowthPath Partners, where she serves high-growth businesses in three distinct roles as a fractional CMO, an executive advisor, and an AI consultant. Liza has held marketing executive roles at Smartsheet, Juniper Networks, and Pure Storage among others. And today we’re going to tap into her ability to use Gen AI to drive strategic development. I know we talked about content, we talked about this. But strategic development, I think it’s such an interesting topic. So Liza, welcome. How are you? And where are you this fine day?

Liza: Hi Drew, thank you for having me. And I am in beautiful Boulder, Colorado. So just enjoying the sunshine here.

Drew: Awesome, Boulder is beautiful. I am quite jealous because that would mean that you’re 45 minutes from a ski slope. So sounds pretty good. Anyway, we’re gonna dive in. In our prep call, you mentioned that your personal mission is to elevate the strategic value of marketing. We can do full stop using AI though. And I’m curious what drove you to believe that the strategic value of marketing needed elevating in the first place?

Liza: I love this question because it cuts to the heart of what I’m passionate about, as you mentioned. I’m in the back nine of my career and I started thinking about my post-operator role, right? You know, I may have one or two more CMO gigs in me. As I look at potentially my post-operator role as someone who could serve on a board, I saw that there were only 41 marketers on Fortune 1000 boards, and less than 3% of board members had marketing experience. So I deeply looked into that and why that is, right? And there is this perception that marketers are technicians rather than strategists, that we are valued for our campaigns, our ads, our emails, and social all these things that are highly, highly visible, and awesome, right? But I think what people don’t see that powers all these things are the great work that we do in deeply understanding markets, deeply understanding customers, competitors, how we position, how we ensure that the products that we have truly meet the market’s needs and that we have something differentiated, that’s sustainable over the long term. And, you know, being a CMO that came from the product marketing, and go-to-market strategy pillar, I have a lot of passion for that work. I have a lot of passion for all of those deep customer insights and insights about the market and the competitors that need to be unearthed so that we can create these beautiful and highly effective campaigns. So that is where my passion for AI really precipitated. Because it elevates that, right? Now, more than ever, beyond content creation, we can use AI to deeply analyze data and understand customer insights, personalize the experiences, personalize content, automate things, and use it as a thought partner, collaborate with it, and then ultimately make ideas better. So that’s a little bit about me and a little bit about what I’m passionate about in my mission.

Drew: I love it. And we’re going to spend a lot of time talking about thought partner and how you’re using AI. There are a couple of things that really struck me when you said that marketers are perceived as tacticians. And I wonder, before we get even into AI, there’s so much pressure on CMOs right now to drive pipeline and revenue, because there’s so much pressure on CEOs because look, there’s a hidden recession in B2B, it’s harder to close a sale than it has ever been before and so the pressure on CMOs to deliver. And I’m wondering if those two things are connected, in that CMOs are being forced to focus on pipeline, that’s the metric that they’re getting measured. Obviously, that ends up being tactical. And if we could pull it back and have a broader perspective on metrics and how we’re measuring them we’d maybe have a broader perspective on the role itself. Am I connecting dots that shouldn’t be connected? Or are you feeling that?

Liza: I think it’s a little bit of a misnomer. Here’s my view on this, we are still navigating this evolution from growth at all costs to sustainable profitability. So when we think about pipeline, pipeline, pipeline, it’s really more about really great quality pipeline, pipeline, and reduction of that sales cycle, right? So in the world of grow at all costs, this is about a land grab, creating as much pipeline as we can, and every single prospect could be considered the right customer, right? In this world of sustainable profitability, it’s the opposite. It’s narrowing the aperture, focusing on maybe two or three segments of the market, personas, or use cases, and going deep in those things, right? Because when we are able to truly hone in and focus, that’s when we can deeply understand them, really craft messages that resonate with those customers, craft a product that addresses their key pain points, right? And when that happens, hey, marketing doesn’t have to deal with peppering their money across a very broad swath of segments. We can now focus on those key ones, sales can now have better quality pipeline, and hopefully shorter sales cycles. So I think this navigating the evolution from growth at all costs to sustainable profitability is a difficult one because they’re polar opposites. We’re now more interested in retention, loyalty, and increased lifetime value versus just growth of pipeline and worth of revenue.

Drew: We should be interested in those things. I think it’s a key thing I want to just make sure is because even in that sustainable profitability, there is still tremendous pressure to drive pipeline. And what you’re saying is, it’s got to be better pipeline, it’s got to meet your ICP, which is now more restrictive, but it’s still about pipeline when it should be about retention and loyalty and lifetime value and other things, even maybe employees.

Okay, but we’re sidetracked because we want to get to the conversation about using generative AI as a thought partner to develop strategy. I had a little bit of time to go through a couple of the documents that we’re sharing in chat, the “10 Steps in 10 Minutes” and “Strategic AI Collaboration Tips with ChatGPT.” It’s wonderfully detailed, and I encourage everybody to go ahead and review that after Liza and I finish our conversation. And for folks that are listening to this as a podcast, we’ll definitely share those documents in the show notes. But let’s talk about this framework because here’s the interesting part of this: in many of the conversations that we’ve had so far about generative AI, there’s a lot of notion that it’s average, you know, AI stands for “average information” or “average intelligence,” because it’s bringing everything out. And when we’re talking about strategy, we’re talking about the opposite because you’re really trying to develop something unique. So talk a little bit about the process that you talked about in the “Strategic AI Collaboration Tips,” and get us to understand how these tools can, in fact, be thought partners.

Liza: I’ve observed a lot of focus on content creation when it comes to AI, right? And I’m so empathetic to the journey that everyone is going through right now. I’ve often said that it is as if OpenAI, in November of 2022, dropped off a bunch of really powerful Legos at our doorstep. And unfortunately, it came in a box that has no picture nor instructions for what to build. We have this little conversation starter box at the bottom, and it’s like, “Start typing,” which is really, really interesting, right? And we’ve never seen technology like this before that is so accessible for the masses. And as we think about the low-hanging fruit, people gravitated towards content creation. We saw people using it for creating social posts or summarizing articles, or helping them with their emails, or even creating images for ads and for social, right? But I believe that we are just scratching the surface when we think about AI from that perspective. And, you know, Drew, you mentioned about using AI as a thought partner, that is just one of the things, right? You know, McKinsey’s report says that the investment in AI in sales and marketing could add 1.4 to $2.6 trillion of business value. Now, I believe that we can’t get $2.6 trillion of business value by just doing content creation, right?

Drew: Correct. No doubt, particularly if the content doesn’t include original IP, you know, and then we’re just iterating and creating average content. So yeah, I mean, the use case of content creation, we’ve covered. Certainly, if you are Nicole Leffer, you can do crazy, great content doing it. But I don’t think the level required to do that is a lot higher than folks are willing to put the energy into. Your other points on the chart: collaboration and ideation, automation, data analysis, insights, personalization, all, to me, the more interesting territories of this. I think I want to make sure that we get back to the strategy portion of this and how to use it that way because it’s an infinite conversation. So let’s keep going past that.

Liza: Yeah. So what you will see today, I actually have four use cases for you that focus more on the strategy piece of this, right? How we’re using it in a collaborative manner to make ideas better and how automation is involved in that, right, and deep analysis of data, right, and getting some key insights. And then ultimately, that strategy leads to personalization. So I’m going to be talking a little bit more about that. But one thing I wanted to just emphasize in this slide is when we think about AI, most people think about productivity and increased efficiencies. But there’s this study that basically says, you know, this is from the Harvard Business School study where they took a bunch of BCG consultants, half of them used AI, the other half did not. 12.2% of those that used AI completed—the ones that used AI, completed 12.2% more tasks, did things 25% faster. And the kicker here is the quality was much higher by 40%. So think about this: in using AI as a thought partner, our ideas, we feed it our initial ideas, we feed it our frameworks, we feed it our hypotheses, and AI can help us in collaboration with it, help us make those ideas better.

Drew: Yeah, first of all, I think about it as BCG consultants are very smart. They know how to do this. And one of the things that I know we’re going to get into, and where content folks tried and kind of went “that was crappy,” is that this is an iterative process where you’re going and going and going, just like you would with people in strategy development. And I think if there’s one big takeaway from this is if you’re just taking the output that you get from the first query, you’re missing the point of all of this. So that’s one big, big takeaway. And it’s so clear when you read your document, you talk about 10 steps because it is 10 steps to get to some really interesting ideas. Okay, so keep going.

Liza: Awesome. By the way, I’m going to go through an example of how exactly I did that: start with a thought, and then keep honing in the thought in collaboration with the AI tool. So I’m going to show you four use cases of fairly strategic endeavors in marketing. But know that there’s an infinite number of use cases that are strategic in marketing. Like here is just a sampling of relevant use cases by function and marketing. This is not even all of it. Like every day, we’re uncovering more things, you know, as we make our teams more effective and efficient. So I’m just going to be picking four of these to show you what’s possible. And that’s exactly what we’re trying to do here. It’s just to inspire what’s possible. And these are the four use cases: I’m going to start out with a competitive analysis and positioning use case, and then we’ll go into segmentation and targeting, really around picking the high-value segments that align with the company’s strengths. And then we’ll go into a competitive defensibility analysis. This is where we look at a company’s position to determine whether it is sustainable or is this something that it can be easily replicated by your closest competitor. We’ve seen a lot of this happen in the world of AI, whenever OpenAI or Google launches something, how many startups close down their business or are devalued as a result, right? And then this last piece is around thought leadership, you know, to what Drew was mentioning earlier, have some ideas that may collaborate with AI to make those ideas better. Okay, so just really quick, before I go to the competitive analysis, know that while we focus a lot on AI tools and the responsible use of those AI tools, your domain expertise, your unique knowledge of the data, and the data that you that are uniquely within your company, those are the key things, you know, just like a hammer can be used to build or destroy, AI is the same thing. It’s just the tool, right? The value is you, really in our domain expertise and in our unique knowledge and data. And then ultimately, hopefully, we’ll get more efficient and better-quality results.

Drew: Yeah, I just want to put a punctuation point in: it’s garbage in, garbage out. It’s the same whether it’s a database when you’re just trying to do a mailing list, if you have a bad list, bad data, domain expertise is everything. So the key insight here is this is your collective insight as an organization, and yours is a strategic driver of your business as the CMO that you’re bringing to any application. And I just think that’s so important. This doesn’t do the thinking for you. It helps take your thinking and make it better.

Liza: I love that punctuation. But let me do like an exclamation point on that one, right? Like make this really sink in: just like we cannot out-exercise a bad diet, we cannot out-campaign a bad product-market fit. So a lot of what we’re going to do is really in that strategic piece. We get that right. AI will elevate and accelerate what’s right. If underneath, it’s wrong, AI will accelerate, and elevate the misalignment between product and market fit, right? So I think it’s just so critically important that we understand that.

Drew: Yeah, I’m gonna add this. So in the 80s, when advertising was the dominant channel for marketers, the expression was there was nothing like a great ad to kill a bad product. And I think that’s exactly what you’re talking about.

Liza: That’s right. This is an overview slide to show you the steps that we took. So my team and I conducted a competitive analysis and positioning. And then I will go through each one of these steps. For the purposes of an example, we’re comparing two project management vendors, or work management vendors, and they’re Asana versus monday.com. And we essentially wanted to understand in which situations would one be better suited than the other, and in which areas is one better than the other based on customer reviews. So for the purposes of the exercise, we can pretend that we’re all in the market for project management software, and we’re looking at evaluating these two vendors. So we go to G2, which is basically a review platform. Lots of customers go in there and provide reviews and they rate the different vendors. So we did a customer review analysis by simply taking screenshots of Asana versus monday.com. And in G2, it basically spits out side-by-side comparisons based on customer reviews. Take those screenshots and put them into ChatGPT. And I’ll show you how we did that. And based on the analysis of ChatGPT, we can gain some insights around how one is better than the other in various dimensions. Additionally, on top of that, in step two, we basically guided it to go look at third-party web resources that are available out there. So from things like buyers’ guides, or analysts’ reports, or blogs, or articles that compare Asana versus monday.com, and give us an analysis based on those resources. And then lastly, on step three, is we combined the analysis outputs of step one and two, and basically said, based on everything that you’ve done and researched, what does it look like? And what does the typical ideal customer profile look like for these companies? So let me show you how we did that. 

So here are the screenshots, right? So you can see compare Asana and monday.com. And in the screenshots, it shows you a distribution of what companies, what types of companies, and what verticals they represent. Also, you could see the different features where one is better than the other, blue is better. And gray is basically second place. And then you can also see some information about the pricing. This is just three of the screenshots, it was actually like eight pages long. But it was very detailed, right? So you can either download the sheet or CSV file or just take screenshots and upload those into ChatGPT. So here’s what we did. In this case, I actually uploaded a PDF of what you just saw, it could be just the screenshots. So I wanted to give you that option, right? And sometimes you just kind of have to play with it, to see what works best. So the prompt was, “I’m a potential buyer of project management software, you’re an expert in this space and have knowledge of various competitors. Attached is a PDF of a summary of customer reviews for Asana and Monday on G2, please review and deeply understand it. What can you tell me about Asana versus Monday based on this information?” So you can see here that it began analyzing all that information from target market to pricing to the different capabilities from task management to project management capabilities, and remote work, and it provided a little bit of a summary. 

So from there, you know, when we go into step two, the prompt was, “Please browse the web, and find any articles, buyer’s guides, analysts’ reports, blogs, social posts, or any third party reference materials that compare Asana versus Monday, please show me the list of resources and the links to the resources you used for the analysis.” So here, it began compiling all of that information and started analyzing based on what it found, you can see these little quotation marks because those are the citations. If you click on those, you will find the sources. So again, you could see that it did some analysis on some key categories for this particular domain. And then I really wanted to see the sources, I didn’t want to have to click on all of those notations. So I said, “Okay, give me the list of these third-party resources,” because I wanted to ensure that these were real, and that it wasn’t hallucinating or anything like that. So it did provide all the resources. And I began clicking on these to ensure that these were actual articles that compare the two vendors, and they were. So at least for this exercise, I didn’t see any hallucinations at this point. 

And then I had guided it to then combine the analysis from steps one and two, that basically says, “Hey, take everything that you know now, and show me a radar chart and show me where one is better than the other based on customer reviews in these eight dimensions.” So you can see here that Asana ranked higher when it came to portfolio management and overall satisfaction, and then Monday ranked higher in the areas of pricing and user sentiment and then collaboration. What was also important here is that it also told us based on customer reviews, what types of customers tend to favor one versus the other. So Asana was ideal for teams and organizations looking for robust portfolio management and extensive integration without user limitations. It talks about the strengths, talks about key segments of the market, so small to medium size, and then did the same thing for Monday, and Monday was ranking higher among larger teams and organizations or businesses across different sizes. So if we think about this as marketers, this should intuitively tell us that our potential customers are going through this kind of evaluation of our products. So if we are not in those articles, in those publications, in those buyers’ guides, or if we’re not represented in the right way, then we’ve got a challenge. We also can look at how the customers are responding to these reviews, and basically go, “Oh, my gosh, we’re being positioned in that way. And we’re not that, we’re not for small teams, we’re actually for large teams, right?” So you can then reposition, and we think about how do I get more larger companies to do reviews? Or how do I get them to think about our AI capabilities, because no one’s talking about our AI capabilities? So I think from both sides, either as a consumer of the product, this is something that they would go through, or as a marketer or a salesperson, our Go-to-Market Strategist thinking about how we position in the market? Is this how we want to come out in the market and how we want customers to perceive us?

Drew: What I’m sensing is, this is taking a cumulative amount of information, and sort of averaging it out and saying, on average, Asana is over here, and on average, Monday is over here. And so if that’s not where you think the brand is and you think “Oh no, Asana is saying we’re for a large company,” then your messaging or your targeting, or your something isn’t getting through. So you can sort of figure out potential problems because this is almost like you did a tracking study. And you’re asking folks, but you didn’t have to pay for it. You just went out and got it.

Drew: Before you keep going, because this is great. I just want to read the ending to your last prompt, which was, “Think about this next step-by-step, take a deep breath, there is no rush, I have full confidence in you. Do you have any questions?” And this is so funny, because when we had Andy Crestodina, he kind of felt like, why are you being polite? And of course, Nicole is a big believer in being polite. And so you’ve obviously embraced the talk to this machine as if it’s a human.

Liza: Yes. Certain things work, and certain things aren’t necessarily necessary, right? I do feel that when you give it the opportunity to think more carefully, take a deep breath, slow down, right, just like human beings, when it’s not rushed, I think it does better in responding. It also does better when you encourage, just like humans, and in fact, I have found, and others have also seen this, that when you tip it, like literally tip, but I don’t mean exchange dollars, when you say, “Hey, I’m gonna tip you $100,” it actually comes up with a better response. And I have proof of that. And it also responds well to sticks. So if I say things like, “This is so important to the advancement in career, if we don’t get this right, there will be severe penalties.” That even works, right? I do know that being polite, it doesn’t necessarily respond better to politeness. However, if these models are being trained on how we interact, I would rather be trained that humans are polite and respectful beings, right? I’m doing whatever I can to ensure responsible use of AI.

Drew: Love it. Well, we already know that empathy is a wonderful trait for a leader often underestimated in the C-suite. But it sounds like a little empathy when you’re using AI is also a benefit. So very cool. Okay, so keep going. So we now have this, is there another step involved here? Are we done with this one now?

Liza: We’re done with this one. Also, notice that I gave it the opportunity to ask questions. And the reason for that is sometimes we’re not good prompters, we forget, you may have forgotten to add an image or didn’t give it enough context. And this is me checking myself that I give you enough so that you can do a good job.

Drew: Last question on this before we go on. So in this particular one, you just used ChatGPT. I know and others, you’ve gone back and forth between Claude and ChatGPT. Are you going to do that in the next example or another example? Or should we just talk about why you do that now?

Liza: We can talk about that in the fourth example, and in the second guide that I sent to you, Drew, I offer up recommendations and steps for how you use both simultaneously.

Drew: That’s great. Okay, so keep going because we’re gonna run out of time. So here we go.

Liza: Alright, so the second use case is really about targeting segmentation and targeting, right? So we basically said in the world of sustainable profitability, you can’t go broad, you got to like narrow in, right? Who are you going to go after and then go deep. I work with a lot of executive teams. When times are tough, the automatic response would be, “Oh, let’s go after more segments, more customers, more use cases, more applications,” that’s not how the world of sustainable profitability works, you got to hone it, right? You got to pick some segments, whether it’s two or three segments to go after that go deep. So this is the exercise that my product marketing team and I use to help align an executive team because the CEO wanted to go after a new vertical, the CFO wanted to go after yet another vertical because there was a lot more opportunities there because it was a growing market, the CPO wanted to go after a different vertical because of a better market fit based on what’s on the roadmap. So it was really difficult to align the C-suite. So what we did was use data and use AI to come up with this. And in this process, we essentially this is a heat map. This is the actual output from ChatGPT. And I redacted all the information, no confidential information or proprietary data was fed into GPT in creating this, but in essence, what we’ve done is we’ve selected our top five segments that we wanted to evaluate, there were actually eight. And then we highlighted key evaluation criteria. That we would go through and evaluate each one of these segments on. What we did was we fed ChatGPT with information around market size and growth based on analysts’ reports like Gartner or Forrester, which had some information about competitive intensity based on data from customer review sites like G2 or Capterra. Our win rates, we fed the data from our win-loss data from our CRM, which was HubSpot. And then we also did a lot of market research in talking with the product team to determine fit with our product, which segments best leverage the product capabilities in our roadmap, our partners’ strength was based on data that we had on partner referrals, leads, and conversions. And then we also have a customer reference database. We did some evaluation across all of these things. And we can do it just straight line without weighting of the different criteria, or we can actually apply weights to the criteria. So we’ll show you how we did that. And at the end of the exercise, it created a heat map to show us the top segments were across all of the evaluation criteria that we should potentially go after. 

Okay, so what I did was we’ve actually created a custom GPT for this. So we didn’t just do ChatGPT straight up the original one was, but now what I’m going to show you is a custom GPT that I’ve crafted so that we can do this over and over again. And in fact, Drew is going to share with you guys the custom GPT link so that you can use it for your environment. So here I’m going to show you how we did this. And I’m just going to show you two of the rows otherwise it’s going to take a long time to do all of the different rows. So let’s just look at the market size and growth. So this is just a little tiny spreadsheet that shows the potential opportunity and market size and growth for five segments, manufacture, healthcare, energy, food and beverage, and retail for the years 2025 to 2030, uploaded that into ChatGPT and said “You are an expert, market researcher, and analyst in the supply chain management space, please review the attached sheet and understand it, analyze it and provide a summary of your key takeaways, then I will give you further instructions.” So here it did some analysis of it and without me asking it to calculate the growth rates, what’s interesting about this is it did. So I didn’t even need to prompt it to calculate the growth rate over that five-year period. And then it had some key takeaways around the highest growth sector where we can achieve more steady growth and it also highlighted key factors in healthcare and manufacturing. Now this is where we had to guide it. 

So now I said, “Fantastic.” So it loves encouragement. Now we basically said, “Please create a table with the industries in the columns in this order.” So I gave it the order. Then the rows are the 2025 market size and then the growth laid across five years. So it then created that table. Looks great. Now turn this into a single heatmap where the biggest market size gets a five because we have five segments. The smallest market size gets a one. Growth rate, highest growth rate, gets a five, lowest gets a one. So here on the right side is the heatmap for those two columns. So if you can imagine we did this for the eight criteria, did the total, and did the full heatmap, as you saw, on the slide, now, we took this into the C-suite. They’re blown away. The value wasn’t in this heat map, the actual value was actually in the conversation, to talk about how we evaluated the stakes that people really saw the segments from a 360-degree perspective, rather than just somebody going, “Oh, I think we need to go after healthcare because it has the largest market opportunity” and somebody else says, “oh, we need to go after retail, because it’s growing the most.” But now they can see all of the different parameters by which we evaluated each one of these segments. Alright, let me pause there to see if there are any questions.

Drew: You threw a lot at us. Without going through and thinking it through, it’s hard to wrap your mind around it, I think my main takeaway, again, is where you concluded, which is that this led to a really strong conversation, and hopefully helped identify where the company should invest, right? And as part of their analysis, obviously, none of this is you’re just going to do what it says, you’re gonna bring your wisdom to it. And I do want to go back because one of the lines on your thing was customer reviews. And we did get a comment from a Huddler, who noticed that use of G2 which some folks would say the quality is kind of low, or it may not be as good as sort of direct feedback. But I also think you show that you went out and asked the machine to go find other sources, not just relying on G2, you really do have to make sure that you have good data to work with to do these analyses. Right. So do you want to address the G2 thing?

Liza: Yeah, so G2 was just one of the data sources that we had, because that’s just customer reviews. We also did customer interviews, we did about 15 of those. So there was a lot of analysis that we had to feed this with customer interviews. We also did stakeholder interviews on this one, as well as partner interviews, because we also have an indirect channel that we use to go after customers, right? So all four of those things in combination kind of led into that competitive aspect.

Drew: Interesting. So we have a couple of follow-up questions. One is, was the data validated by internal teams?

Liza: Yes. Let me tell you more about that. So we came up with an answer after the discussion with the C-suite. But we basically said that those are our hypotheses. Now we have hypotheses that we need to validate into the market. This shouldn’t be, “Hey, whatever ChatGPT spits out is what we need to do.”

Drew: And one of the things that ChatGPT is really good about, we talked about this before, is finding the gaps. And so you’re creating hypotheses. And you probably could come up with most of these on your own, but it’s going to find all of them. And it’s going to help you sort of sort through them. The way our brains work is something may be obvious to us. So we’ll just forget it. It’ll also bring the obvious thing, just in case to sort of make sure, so I think that’s important. One question we had is do you have any prompts or easy paths to generating a TAM in ChatGPT? Just do all the work here.

Liza: Oh, my gosh, no, I haven’t figured out how ChatGPT can actually create the TAM. We would have to feed it a lot of information, just like we would have to pretend that we’re the analyst firm that’s creating the TAM. What I’ve actually also tried to figure out is, you know how those analyst reports have like a bar chart? It only shows us one number, but it doesn’t show us the numbers for each one of the bars. So I’m like, “Hmm, I wonder if ChatGPT can actually extrapolate and figure out what those other bar numbers would be.” It’s not as good, right? So don’t, I mean, you can try it but don’t hang your hat on it, because it looks at pixel sizes and things like that. And it is not a very accurate representation of what you see. You and I could do a much better job of guessing what those numbers are than what ChatGPT would do.

Drew: Right. Yeah, I mean, in general when it comes to math, you really don’t want to use that, that’s what a calculator is for. That’s what you know, Excel is really good at and other things but think about what ChatGPT is really good at, and it’s not necessarily math. One more question then we will probably just jump to the fourth example where you did the dueling ones, but since we talked about customer interviews, how have you used ChatGPT to analyze these interviews?

Liza: Yeah. So we would feed it—Here’s what we did, so the interview probably has a set of 10 questions. And then we would have a template that says, “Fill out this template or a table that highlights for us in general: old way how customers used to do it, new way, how they’re doing it today, and what they say about the impact of using us.” So now we feed it all this data, it goes through all those questions, and it begins to populate a table that shows us old way, new way, impact. Well, guess what? Now it’s great. It’s something for me that I can use for messaging, for content creation, for infographics, and all sorts of things. But the key there is somebody, us, had to think about “What do I want to extract out of this thing? I wanted old way, new way, impact” that’s quantifiable, right? Then only then can we start creating content. Does that make sense, Drew?

Drew: It SO makes sense. And it’s so important. And I will share, we’re working on a really difficult challenge of creating a metrics, sort of a new way of thinking about metrics for CMOs, and we had seven folks from the community join us in a task force. And we recorded all the sessions. And I fed the interviews into Claude because at the time, ChatGPT couldn’t actually do it. But I only asked him for quotes. And I got quotes, and it was fine. And I even had to put parameters on what kind of quotes it was looking for. But what I didn’t get, because I didn’t think to ask about is, let’s sort of create some boxes in a grid to give it specific instructions. So I think that’s so important. We are running out of time, and I want to make sure that we wrap things up in a really positive way.

Drew: So I have one question for you. And then I have another question for you. But let’s just talk about the going back and forth between ChatGPT and Claude, and making them two, having two thought partners instead of one.

Liza: So it’s not necessarily just having two thought partners and I will address that. But it’s also in helping minimize or mitigate hallucinations. Talking with two partners or two AI tools, right? So I might ask that, let’s say I’ve got a thought leadership article that I’m working on. And I ask ChatGPT, “What do you think about that?” Then it tells me something, right? “Oh, I think these are the good things about it, I would recommend that you do the following.” I’m like, “Wait, awesome.” I feed the exact same instructions to Claude, or now Google Gemini, right? And I said, “Alright, here’s what I’m doing. I give it context, then this is what ChatGPT said, what do you think of it?” Right, it gives me a response back and guess what, oftentimes, they actually collaborate really well. And they support each other, right? But sometimes they would say things like, “Oh, but here are a couple of other things that you might consider.” Or I’ve actually seen it said, “I have a different point of view.” Right. So now instead of Liza just having a conversation with me, myself, and I, I now have a strategic team of five or more because of the power of AI, right? And then it mitigates hallucinations because if one happens to hallucinate, the other one will catch it. And God forbid that they hallucinate at the same time. So I think there are two key things for me. One is mitigating hallucinations. And then secondly, making the idea so much better. Because you can feed one the response from the other and get them to collaborate with each other.

Drew: And they can find holes. And this is another thing where the holes in this argument and I encourage everybody to take a look at the five steps and five minutes strategic AI collaboration tips for both ChatGPT and Claude because you show the prompts going back and forth in that and it’s so helpful to see it in, it’s worth going through this process and almost replicating exactly what Liza did so that you can see it and go through it. So I think that’s a key takeaway there. Also, we have shared your two assistants, the business prioritization assistant, and the competitive defensibility analyzer. Talk about the difference between those two GPTs.

Liza: Alright, so the business prioritization assistant actually lines up with what you saw here because now you can do this yourself. Now it won’t ask you for all the data and all that stuff but it will essentially allow you to go through the process, right? So here, it’s saying, hey, how many segments do you want to evaluate? I said, eight, and then it’ll ask you what the segments are, right? And then it’ll ask you what the evaluation criteria are. So you just populate all that. And then here, they will ask you for your forced ranking. So for each of the evaluation criteria, you need to feed it, your forced ranking, and then it will help you create your table with the forced ranking and identify the top scores. So in this case, manufacturing, energy, and transportation was my top three, and then it will give you the ability to weigh the evaluation criteria, right, if you want to assign weights. And then here, you can see here that the number two and number three flipped as a result of the weighting. So it just makes the process a little bit easier for you. Now, you still have to do all the things that I mentioned earlier about the data and helping it analyze each row. But once you get to the forced ranking piece, this can help you with that.

Drew: And I think an important part of this thinking, folks, is that strategic planning and strategic analysis takes time. And there are a lot of steps that you would go through naturally, whether if you were doing this without the help of a thought partner or two. And so thinking about using these tools in that way, is really a pause, a step back and saying, oh, it’s not an input-output. It’s an iterative process that you have to just keep going and pushing and never accepting the first output that you get, right or even from the source.

Drew: Okay, quickly, what’s the second GPT?

Liza: The second GPT, I’m going to take the rest of the time on this one. So basically, this is to help you figure out whether your moats, your differentiators are defensible in the market, and whether the value that you add is the same or different than your competitors. So here, there’s two axes here, value that you deliver, is it the same as your competitors? Or is it an order of magnitude more than your competitors? And then on the other axis, the x-axis is, are your moats the same and easy to replicate? Or are they super hard? If there’s really hard to replicate, meaning that they’d have to buy your company, they’d have to fire your people, those kinds of things, then you got something unique and sustainable. So this competitive defensibility analyzer asks you a number of questions and then determines where you belong and where you are. Are you a category maker? Are you just the next release, where perhaps the next version of OpenAI or the next release of Google would deem you less relevant to the market? 

So I’m just going to show you really quickly here. So here, it will ask you what category you’re in. So it says employee experience solution, whatever your category might be? And then it’s asking, hey, in the value that you offer, is it the same incremental or an order of magnitude better than your competitors? And in this case, I said, it’s incremental improvement. And then it will ask you, what are your top three differentiators? So here are just a list of potential options, but you could put whatever you want in there that you deem are most important. And in this case, I said, end-to-end workflows, partner ecosystem, and technology. And then from that, it will say, hey, are those three differentiators weighted the same? Or do you want to apply weights to these things? And I said, you know what, I’m gonna put 30% on end-to-end workflows, 40% on partners, and then 30% on technology. And then it will ask you for each one of those three, is it easy, moderate, or hard to replicate? And I basically said, easy for workflows, moderate for partner ecosystem, and then easy for technology, because we’re all building on top of the same large language model, just as an example, right? It does some calculations, not really complicated calculations, you actually see it here. It gives you a defensibility score, meaning how difficult will it be to replicate, and then a value score meaning how much value that you provide your customers compared to your competitors and based on the responses that you gave it, it spits out an answer. Here, you can see that you are just a “Next release.” So this requires the executive team to really internalize and be introspective about its position in the market and talk about how we go up and over. Not everybody can be category maker. But how do we make our position more sustainable in the market?

Drew: Great. This is amazing. You’re a B2B CMO, and you want to use LLMs to help you develop strategies, give us two do’s and one don’t, for using LLMs as strategic thought partners.

Liza: I would say that use the LLMs to constantly take a pulse of the market and ensure product market fit. I showed you three examples for how to make sure that we’re ensuring product market fit. And then what I would say on the don’t, and I don’t like “don’t”, don’t assume that this is just like any other inflection point that we’ve gone through. This is one where we need to get our hands on it. Regardless of whether you’re the specialist in marketing, or the CMO, or the CEO, we can’t understand the full value of AI until we get our hands on it. So that would be my recommendation.

Drew: I love it. Well, Liza Adams, thank you so much for joining us today. A big Huddler thanks to that. We will share your LinkedIn profile. Is there anywhere else that folks should find you?

Liza: Yeah, you could also go to my website, which is growthpath.net. And you could see a lot of my thinking there where I focus and some of the frameworks that you saw today.

Drew: Awesome. Thank you so much. 

If you’re a B2B CMO, and you want to hear more conversations like this one, find out if you qualify to join our community of sharing, caring, and daring CMOs at CMOhuddles.com.

Show Credits

Renegade Marketers Unite is written and directed by Drew Neisser. Hey, that’s me. This show is produced by Melissa Caffrey, Laura Parkyn, and Ishar Cuevas. The music is by the amazing Burns Twins and the intro VoiceOver is Linda Cornelius. To find the transcripts of all episodes, suggest future guests and learn more about CMO Huddles or my CMO coaching service, please visit renegademarketing.com. I’m your host Drew Neisser. Until next time, keep those renegade marketing caps on and strong!