Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
Fixing Business Bottlenecks with AI to Go Further, Faster w/ Liza Adams
404 Plays2 months ago
In this episode of AI-Driven Marketer, host Dan Sanchez speaks with Liza Adams, an AI Advisor and Fractional CMO at GrowthPath Partners. Liza shares how AI can help businesses identify and resolve operational bottlenecks to go further, faster. From her experience advising CMOs to leveraging AI for real-world challenges, Liza breaks down how marketers can use AI to streamline productivity, improve output quality, and foster innovation. Tune in to hear practical AI-driven solutions that will elevate your marketing and business strategy.
Resources Mentioned:
Timestamps:
- 00:01 – Introduction to Liza Adams and the topic of using AI to fix business bottlenecks
- 01:20 – Common bottlenecks in businesses and how AI can help
- 03:42 – Tackling productivity and quality challenges with AI
- 06:59 – Case study: Using AI to focus on key market segments
- 11:31 – How to leverage AI for market sizing and forced ranking
- 14:45 – The importance of executive alignment and validation
- 20:41 – Using AI to develop campaign strategies and improve messaging
- 26:00 – Creating custom GPTs and enhancing sales processes with AI
- 32:04 – Leveraging customer reviews and data for more effective marketing messages
- 41:00 – Validating AI outputs and ensuring quality with AI tools
- 44:00 – Final thoughts and where to find Liza Adams online
Recommended
Transcript
Intro
Introduction and Guest Overview
00:00:05
danchez
Welcome back to the AI Driven Marketer. I'm Dan Sánchez. My friends call me Dan Sánchez, and we are on a journey in 2024 to master AI so we can make the most of marketing because we all know that marketing and always has more to do. So with AI, we're hoping to leverage it so that we can do more with less in order to catch up. Today, I'm excited to talk to Lisa Adams, who is going to be helping us break down our bottlenecks and using AI to fix the things that are slowing us down in marketing. So welcome to the show.
00:00:37
Liza Adams
Thanks, i'm I'm so excited to be here. Thank you for having me.
00:00:41
danchez
Absolutely. i Someone recommended you on LinkedIn and after checking out your feed and what you were talking about, I was really impressed. I'm like, oh my gosh, got to have her on the show. After we had our pre-interview, I was like even more impressed.
00:00:53
danchez
I was like, no wonder why you're speaking at MayCon. I'm like she yeah you have the goods and I'm excited about this episode and ready to get started.
00:01:02
Liza Adams
Thank you so much.
00:01:02
danchez
So it's okay with you.
00:01:03
Liza Adams
I do learn from many people and I try to apply them where I can.
Addressing Marketing Challenges with AI
00:01:09
Liza Adams
And, you know, I love the AI community with a lot of sharing, we're able to use AI responsibly and and learn from each other.
00:01:18
danchez
So if it's okay with you, let's dive into the deep end and see where it goes.
00:01:20
Liza Adams
Yeah.
00:01:23
danchez
I'd love to know, like, as you're advising, consulting a lot of different companies, as as a fractional CMO, as an AI advisor, what are the kind of issues you're usually coming up against?
00:01:34
danchez
Are they usually marketing strategy issues, tactical issues, tech issues? Like what kind of bottlenecks are you usually coming across as an advisor?
00:01:42
Liza Adams
Yeah, so my work spans fractional CMO work. So with fractional CMO work, you are essentially the CMO of the business, right? So you are in charge of everything from strategy all the way to execution and leading that team and ultimately representing that team to the board. On the other end of the spectrum, I am an AI advisor. So as an AI advisor, I advise CMOs and their marketing teams. And that could be
00:02:13
Liza Adams
a wide gamut of of challenges, right? It could be very simple as inspire our teams with what's possible with AI.
Scaling AI in Marketing Strategies
00:02:22
Liza Adams
So, you know, that might be, hey, we're simply using it for content creation to create blogs or summarize reports. Is that it, Lisa? You know, is that all we we we should think AI for? To help us on this journey, we need to scale with AI, and we need to go from exploration to integration, how do we do that as a team? And how do we ultimately not just do this for the sake of marketing, but also for the sake of the business? Because we are the growth engine for the business. So I see a wide range of of challenges that AI can address. And as a fractional CMO, it keeps me real.
00:03:08
Liza Adams
Because the the use cases that I tend to talk about on LinkedIn or in in podcasts like this one, they're real, right? We're addressing true challenges that marketers and and businesses are going through. And then we take those best practices and and takes a lot of that experience and apply it into my advising workshops and roles.
00:03:32
danchez
It's interesting when I'm talking to people, it is a pretty common question of like, well, what, what can I be doing with this? And usually I'm like, well, where are you stuck right now? And then we'll start there. It's kind of like AI is this general tool that can cover so many different problems.
00:03:41
Liza Adams
That's right, that's right.
00:03:46
danchez
So you might as well just start with the problem. That's slowing down the most, right? Cause that's where you're going to get the most bang for your buck out of it.
Productivity and Quality Improvements with AI
00:03:52
danchez
Like, uh, if we can fix this one thing, how would that change your business? Right. Is kind of how it goes.
00:03:57
Liza Adams
That's right, that's right, yeah.
00:04:00
danchez
Or do you see any trends as far as what businesses are running into the most? Or is it just kind of depend, it's different from business to business?
00:04:05
Liza Adams
It's different by business. However, the natural inclination is to go straight into productivity. like How do we get more things done or how do we make this thing go faster?
00:04:20
danchez
Yeah.
00:04:20
Liza Adams
My team is overburdened.
00:04:22
Liza Adams
Is there a way for AI to help and take on some of these mundane and repetitive tasks? So there's a group of use cases or and a set of problems that tend to address productivity challenges. And then there's another set of problems that tend to address, we need better quality.
00:04:43
Liza Adams
Because of lack of time, lack of resources, we don't have the bandwidth nor the time to really think deeply. right you know When we do research, when we have to do interviews, we don't have time to do all of the data analysis. We don't have the tools to do that. We need new insights to inform decision making. So just better outputs.
00:05:07
Liza Adams
And then the last one is that I see less of this because there's so much focus on you know the productivity aspect of it and then making things better. I do hope that more will start to lean in on making things different. So now we're in the innovation aspect of this.
00:05:27
Liza Adams
It's not just making what we currently have better in terms of faster or better quality. It's not thinking about things in a very different way. So those are the three buckets, if you might, you know and that that you could think about it in.
Case Studies on AI in Marketing
00:05:44
Liza Adams
It's like faster, better, and then different or innovative.
00:05:49
danchez
Yeah. It seems like that's a natural progression.
00:05:51
Liza Adams
All right.
00:05:52
danchez
Like you'd probably, if you're starting to get AI and train your whole team in it, naturally the easiest thing to do is just to be able to do things faster.
00:06:00
Liza Adams
Master, that's right.
00:06:01
danchez
And then of course, as you get better at dealing with AI, and I think everybody goes through this cycle with AI where they actually prompt enough to begin knowing how to prompt in a way that you can get better outcomes than you would have gotten naturally, not only the faster, but you move quick, you get better. And then, and then of course, once you really start to master the faster and the better, you start to think of like,
00:06:21
danchez
I don't know, you start getting eyes to see like, Oh, I wonder if, and you start asking bigger questions of like, now this whole new thing is possible.
00:06:21
Liza Adams
so
00:06:30
danchez
or We couldn't have done before, but you wouldn't have got there unless you had learned how to go faster, gotten better.
00:06:31
Liza Adams
that's right
00:06:36
danchez
And then that's kind of the training you need in order to be innovative with it.
00:06:39
Liza Adams
That's right. That's right. That's right.
00:06:41
danchez
What are some case studies that you're proud of that you've addressed with teams now that you've seen all the way through you guys found the problem, uh, found how AI can become a solution to the problem and implemented it and it worked remarkably well.
00:06:53
Liza Adams
Yeah, I'm going to share with you a use case that we did, my team and I did when I was a fractional CMO for a company that had significant growth during COVID.
00:07:06
Liza Adams
because a lot of their customers went, obviously many of us you know started working from home, and their business was one of those that really thrived during COVID because it enabled remote capabilities in face of that nature.
00:07:22
Liza Adams
But post COVID, it was really tough. right And the company started to think about, what do we need to do to to grow this business, to identify more opportunities? And the natural inclination of the business was to expand into new markets. Go into more segments. Identify new personas and pursue them. But the challenge is when you start going into more markets, you now start to spread yourself so thin.
00:07:52
Liza Adams
And you risk not being able to serve all those markets super well. There's a danger in that, right? You spread your resources, you spread your budget so thin that you can't serve any one or two well.
00:08:04
danchez
Oh, yeah. Yeah.
00:08:09
Liza Adams
And there's a risk of you know yeah increased risk of churn.
00:08:09
danchez
Yeah.
00:08:17
danchez
Yeah.
00:08:17
Liza Adams
so
00:08:18
danchez
Yeah.
00:08:18
Liza Adams
the The idea really is to not expand into new segments. It's to pick your top two or three segments that you can serve the best because nobody else can do what you do. You have exceptional product market fit. And because you're so good at it, you could do it profitably. So we actually use AI.
00:08:43
Liza Adams
to determine which top segments of the market we should go after, rather than identifying new market segments in addition to what we have. right We were actually narrowing the number of segments that that we should pursue. So we we used AI by I've explained this in in one of the newsletters that I launched
AI in Market Segment Evaluation
00:09:06
Liza Adams
in just the last couple of weeks. You know, we have this framework by which we evaluated the segments across a number of criteria. It could be market growth, market size, competitive intensity, strength of partnerships, a Fitwood roadmap, and evaluated each one of those segments by each one of the criteria.
00:09:31
danchez
So you came up with all these segments. Did you give each one a score of how to rank them?
00:09:34
Liza Adams
That's right.
00:09:35
danchez
Okay. Did AI help you come up with the segments and the scoring factors?
00:09:35
Liza Adams
know but So we came up with the segments because we had an idea of which segments we wanted.
00:09:41
danchez
Okay.
00:09:45
Liza Adams
And let's just say there were eight of them.
00:09:45
danchez
Yeah.
00:09:47
danchez
Yep.
00:09:47
Liza Adams
And then we had our criteria.
00:09:50
Liza Adams
And then what we did, let's let's just say the market size was one of the criteria. We used market sizing data, you know analyst reports data. to to do that evaluation.
00:10:02
Liza Adams
And we did a forced ranking scenario with chat GPT. So basically fed non-sensitive data, we had to redact some of the market sizing information, fed it in into chat GPT, and we asked it to do a forced ranking.
00:10:18
Liza Adams
So largest market size gets an eight because we were evaluating eight market segments, lowest market size gets a one.
00:10:20
danchez
Okay. Yep. Yep.
00:10:25
Liza Adams
And ultimately, you we asked it to do a heat map. So the eight got a red and and the one got a blue kind of thing.
00:10:33
danchez
Gotcha.
00:10:34
Liza Adams
And you kind of do that same thing for each of the criteria, right? You did that for market growth.
00:10:38
danchez
Okay, so lowest score wins then.
00:10:40
Liza Adams
Yep.
00:10:41
danchez
Yep.
00:10:41
Liza Adams
And then ultimately, at the bottom, it had a total, which allowed us to to see across multiple criteria which one was the best.
00:10:45
danchez
Yep.
00:10:52
Liza Adams
So let me just pause there just to make sure that that was landing with you.
00:10:56
danchez
Yeah, no, so I'm, did you have, I'm seeing it. You put, you just uploaded a chat GPT. These reports redacted some information. Probably you probably upload them as PDF.
00:11:06
Liza Adams
Yep.
00:11:07
danchez
then you just write like one long mega prompt that said like, okay, here's what we're doing. Here's what we want to know. Here's the context of the information. Here's the criteria.
00:11:17
danchez
Here's the categories, the criteria and the, and the segments.
00:11:23
Liza Adams
Nope.
00:11:23
danchez
Now go and do it and just get report report back or to do in multiple steps.
00:11:27
Liza Adams
Actually not one mega prompt. It's multiple little prompts.
00:11:30
danchez
Yeah. Okay. So you chain prompted it to go step by step and to think through it.
00:11:32
Liza Adams
Right. Step by step.
00:11:35
danchez
Yep. Yeah.
00:11:35
Liza Adams
And, you know, there's value in going step by step and, and, and smaller bite size prompts, right? Because you can, uh, course correct if it doesn't understand.
00:11:42
danchez
Yep.
00:11:46
Liza Adams
it is also a good way to check AI's work because you're only checking little things versus checking a whole heck of a lot of things.
00:11:46
danchez
Yep.
00:11:52
danchez
yeah
00:11:54
Liza Adams
Right.
00:11:54
danchez
Yep.
00:11:55
Liza Adams
So definitely step by step. Like for for example, that first prompt was I'm going to give you the market size for these eight segments. read this document or read this Excel spreadsheet, understand it well. Then I will give you further instructions, okay? First prompt. Second prompt. Now that you understand all of that, do a forced ranking based on market size and show that to me in a table. but And I told it, you know, I want the segments on the column and the market size is the row.
00:12:34
Liza Adams
So it gave me basically a row where with the forced ranking. Then I said, all right, color code this thing and create a heat map just for that row.
00:12:46
Liza Adams
Then it just did that, right? So now we have one. And I said, now I'm going to go to the second criteria.
00:12:54
danchez
Yeah.
00:12:54
Liza Adams
Now we're going to go to market growth and then did the same process.
00:12:58
danchez
Okay.
00:12:58
Liza Adams
After eight criteria, I said put them all together. Put all the rows together in a table. So now we have a table, a matrix that is color coded based on all the forced rankings. And then I said give me a total with weights that are equal across the different criteria. So now you have at the bottom a total.
00:13:20
Liza Adams
an aggregation and you can see which one won out across all of the different criteria. You can also say, oh, by the way, I want to apply different weights because it's more important to me, market growth is more important to me than a number of reference customers, right?
Aligning Company Strategy with AI
00:13:38
Liza Adams
So you can actually play with the weights and then it will give you a different result.
00:13:38
danchez
Yeah.
00:13:43
Liza Adams
So let me pause there.
00:13:45
danchez
This is such a fantastic process and I feel like is one of those things that probably need to be shouted from the rooftops because I feel like the decay or the the killer of many businesses is operational complexity.
00:13:59
danchez
right because line i mean you You're successful on one thing, but naturally line extension comes in. you're You're serving way too many products to way too many markets and it slows down and kills many companies, which is why when consultants and fractional CMOs like yourself get in, one of the first things we see is kind of like, hey, like the easiest way to profit is just by eliminating work that you're currently doing.
00:14:22
danchez
Not even launching new stuff, just stop doing the stuff that's killing your profit and you win. But this actually gives people a way to... I don't know about you, but like I've recommended that many times that people are like, well, yeah, but we need more revenue.
00:14:36
danchez
yeah like You don't understand by doing less, serving less people as is the path to revenue. And you this process is essentially given others a way to prove it empirically, or at least give show like, hey, no, like it's not just me guessing.
00:14:42
Liza Adams
That's right.
00:14:51
danchez
It's like, here's a path you can take that proves empirically that there are better segments. And if you focus on these segments, we can grow it. So once you're able to kind of prove that these segments are better than these segments, what steps do you take?
00:15:06
Liza Adams
Yeah. it To your point, the reason we did it is, you know, we could decide to add more segments and and we would now peanut butter. And we were just not convinced that we were going to get to the to the revenue goals, right? So we said, hey, there's no harm trying this. Let's narrow narrow the aperture and figure it out.
00:15:28
Liza Adams
But that was only the first step, Dan. you know That was product marketing, doing our little table, you know using data to to craft this table. We needed to get alignment. right We can't just say, hey, this is what chat GPT created for us.
00:15:42
Liza Adams
Here's the heat map.
00:15:42
danchez
Yeah, yeah.
00:15:43
Liza Adams
Therefore, these top three segments are the segments we're going to go after.
00:15:45
danchez
Yep.
00:15:47
Liza Adams
we We went to the executive team, right? Because the executive team had different views on which segments were the top ones.
00:15:54
danchez
Yeah.
00:15:54
Liza Adams
And you know their hypotheses may have been right in one or two dimensions, but not right across eight dimensions. Now that they saw the matrix, let's have a conversation, right?
00:16:03
danchez
Yeah.
00:16:07
Liza Adams
Let's have a conversation around the different weights, the different forced rankings. you know Let's debate these things. And that is where I would argue that that was the most valuable part of the process.
00:16:22
danchez
Yeah.
00:16:23
Liza Adams
the The framework was foundational for them to have a conversation, but that debate within the executive team to ultimately drive some alignment was the most valuable thing in there.
00:16:35
Liza Adams
But even after that debate, you know, healthy debate, let's say we came up with like the top three, the the next, phase of the process was that, all right, let's tell everybody in the company that these are our top three and let's pivot to these three.
00:16:50
Liza Adams
No, we took another two to three weeks and said, let's validate our hypotheses and our data with the market, right?
00:16:56
danchez
There we go.
00:16:59
Liza Adams
Let's talk to stakeholders, to partners. Let's ensure that we got the competitive information, right? Let's talk with customers and prospects. did some of that work, and then ultimately came back again as an executive team, changed some things based on what happened in the validation process, and then ultimately aligned.
Developing Campaign Strategies with AI
00:17:22
Liza Adams
And then once we get alignment, I would you know i would argue it it was probably one of the most strategic things that marketing can do.
00:17:31
Liza Adams
because it avoided, you know, the not a whole company from CEO to frontline managers know exactly the segments that we're going to go after. If somebody has a wild idea about another segments, no, we all aligned on these three segments, right? And then from a marketing perspective, we knew exactly how to allocate our budget and where to allocate it. These are the three segments that we're gonna go after. This is our ICP. We identify personas within these segments and we will craft campaigns specifically for these segments because they are the ones that we can serve the best from the criteria that we chose. It makes it so much easier for marketing to prove its worth rather than having to guess or having to peanut butter its money across the board.
00:18:20
danchez
I feel like marketing always has good insights to what products to go with or kill. But I usually find that marketing is not usually the one in the executive suite to actually make choices around that. And I don't know, I think it's just because there's probably just not enough empirical evidence that marketing brings. But with this process, it actually makes it possible to bring the empirical evidence you need to drive the conversation in a really productive and profitable way. So I love that you're using AI in order to make this process.
00:18:47
danchez
simpler, because certainly you could do it by hand. But gosh, it's a lot of work to create that one report. Doing this manually is exhausting.
00:18:52
Liza Adams
yeah but's right Yeah, I have a blog post on it, but the newsletter goes more specifically into the process that i and we went through.
00:18:54
danchez
But AI is actually able to go through read it and actually do analysis and explain why so that you know, you can double check it to make sure it actually gives you a good report. And I'll make sure to link to you said this was a newsletter. I think you have a blog post on this now.
00:19:09
danchez
Okay.
00:19:14
danchez
Well, I'll share a link to this case study. If anybody wants to break it down and replicate it, I know I will be because this is such a common problem I've run into. And I'd love to know exactly step by step.
00:19:24
danchez
We just heard it, but I'm like, yeah but I want, I'm going to need to reference it later. So I'll be linking back to that.
00:19:29
Liza Adams
Yeah.
00:19:30
danchez
What else have you found AI to be useful for?
00:19:32
Liza Adams
Yeah. So, so once we got down to the three segments, it got super easy. cause now we have alignment. So we actually used, AI to develop, our campaign strategy.
00:19:45
Liza Adams
So you know if if we but let's just pick one of the segments, right? you know we We said, all right, for for this segment, help us brainstorm with us. So we used it as a thought partner.
00:19:58
Liza Adams
And in that thought partnership, there's there's there's a collaboration. you know It needed to understand our market. It needed to understand our positioning, our messaging.
00:20:08
Liza Adams
It needed to understand our personas. So all of that work that we did, we we had to feed into AI.
00:20:16
danchez
yeah
00:20:16
Liza Adams
Again, redacting it and taking out sensitive information. People always say, hey, you're feeding it competitive information and messaging. I'm like, they would find it on our website anyway.
00:20:26
danchez
Yeah. Yeah.
00:20:27
Liza Adams
And in fact, if AI wants to learn that that's our positioning and that's our messaging and that's our value prop awesome, you know, feed it to the market. Right. So. hey You know, I understand that the competitors might see that, but they will anyway, because that's what they will see on websites or on customer review sites.
00:20:43
danchez
Yeah. Yeah. Yeah, yeah.
00:20:46
danchez
yeah
00:20:47
Liza Adams
So feeding that that information around your market strategy, your positioning, your messaging, your value props, your personas is so critical in helping AI understand, giving it context of what you want it to do, right?
00:20:50
danchez
Yeah.
00:21:01
danchez
Yeah. Yeah.
00:21:04
Liza Adams
And then really thinking about, all right, AI, let's brainstorm on what the campaigns might look like in support of this the segment. And here are the three goals that we're trying to achieve. And given all that context, it begins to give you some ideas.
00:21:26
Liza Adams
i and Sometimes it's not quite right, but then you guide it to to give you the ideas. And we actually came up with three campaigns. One of them was one that we never thought about that came from AI. From from that discussion, from that, we used it as a thought partner in essence. it This is where I feel like AI made it better.
00:21:53
Liza Adams
You know, that the first process, it sped it up, right? But it also made it better. This one sped it up, made it better again. And then ultimately, once we have that idea, now we use AI again. All right. Help us to.
00:22:10
Liza Adams
further craft this campaign. What does it look like? Here's the buyer's journey.
Strategic Documentation and AI
00:22:15
Liza Adams
what does How would you design this campaign? So what AI does for us, Dan, is it helps us get started quickly, right?
00:22:23
danchez
yeah
00:22:25
Liza Adams
It like gives us a first draft. Once you give it a lot of context, a lot of really good insights and information, good first draft, and oftentimes that's the hard part. like just getting it lifted off the ground.
00:22:37
Liza Adams
So rather than lifting it off the ground, now we're lifting it from our knees because AI has help helped us to lift it off the ground a little bit, right? And then now what we need to do, it's easier to edit than to start from scratch.
00:22:45
danchez
Yeah,
00:22:49
danchez
always.
00:22:49
Liza Adams
And AI allows us to not start from scratch.
00:22:54
danchez
I find that the strategic documents are actually useful for the first time. How many times have we written marketing plans or persona guides or positioning statements that just sit in a folder in Google drive somewhere, be seen once and never again.
00:23:07
Liza Adams
yeah
00:23:10
danchez
Right? Until we're like, who don't don't we have some personas? Oh yeah, we did that as an exercise eight months ago. You're like, oh gosh, but nobody's looked at it since. AI keeps it front and center though. Do you use custom GPTs to attach it to or do you just upload it most of the time?
00:23:24
Liza Adams
No, we' we're starting to use a lot of custom GPTs.
00:23:27
danchez
yeah
00:23:27
Liza Adams
So thank you for asking about that. We actually did a custom GPT for the strategic one that I talked about. where we created a heat map because we wanted to be able to use the custom GPT when the executives were having a debate.
00:23:46
Liza Adams
While they were having a debate and debate, the forced rankings were changing. So we were doing it in AI in the custom GPT as they were changing the forced rankings, right?
00:23:52
danchez
Yeah. Yeah.
00:23:56
Liza Adams
So the the executives were seeing how the matrix was changing. So that's one custom GPT.
00:24:01
danchez
Yeah.
00:24:04
Liza Adams
The other custom GPTs are a bit more tactical, you know, like for copywriting, for, you know,
00:24:09
danchez
Yeah.
00:24:14
Liza Adams
like to your point, like copywriting, you put brand tone and voice in there. You put persona information in there, right?
00:24:20
danchez
yeah Yeah.
00:24:23
Liza Adams
Okay. Hey, we have this persona Dan. We're going to do an email for him. This is the goal. And then we don't need to explain what the Dan persona is because it's in the knowledge, right?
00:24:34
danchez
Yeah, yeah, it's already in there.
00:24:38
Liza Adams
Here's the goal, and I need three emails to Dan at the top, middle, and bottom of the funnel. Please draft it. Now again, lifting it off the ground for us, all we need to do is edit from there.
00:24:53
danchez
Yeah, I love loading. I just did this for a client just yesterday and I loaded it. I actually did an intake interview with this client just to ask them all the basic information. They were like, what what are you asking? I'm like, what industry are you in? Who's your target audience? What's your story? What's your mission? What are the brand elements that make you weird?
00:25:11
danchez
Like just, I probably asked 50 questions, unloaded it all into an Excel sheet. And then I upload it all to a custom GPT with some other basic information about their business and their products and different things. And now I just query that all the time. It's a great starting point because like you said, and now I can go and take things like, Hey, I love this email from a totally different industry email.
00:25:30
danchez
But I want one like it in this client's industry. But use this as an example. It's like you constantly have all the strategic docs and everything. everything Then you could take inspiration from totally other places and be like, hey, mash these two together. And it just it just does wonderfully with that.
00:25:46
Liza Adams
I love that because then I've said that AI will force us to be more strategic and it will for us force us to be more authentically human, right?
00:25:53
danchez
Yeah.
00:25:58
Liza Adams
This example that you're just talking about, it's forcing us to really be strategic because absent those strategic documents, Well, AI is just going to give you average information. You're just going to be creating content that's a result of what it can find on the web. It's not truly personalized and and relevant to the market that you're going after that you know you can serve the best because of who you are.
Bridging Strategy and Tactics with AI
00:26:27
Liza Adams
So I think it's such a gift to marketers. It will highlight our strategic value.
00:26:31
danchez
Yeah.
00:26:32
Liza Adams
And if we're not doing that, it will force us to be strategic.
00:26:38
danchez
I love that it gives away for strategy to actually be utilized, because you often find that there's a diff there's a disconnect between strategy and tactics and this actually bridges the gap in a way, because you don't have to depend on just people's memory of the strategy now. um'm I'm pretty sure like will I won't force people, but like I'll just be like, hey, always start with this custom GPT when writing anything.
00:27:01
danchez
That way I know all the objectives are taken into account each time, at least by AI, because AI will always remember, right?
00:27:07
Liza Adams
Yes.
00:27:07
danchez
I mean, it it'll hallucinate sometimes, but as long as you're giving crystal clear context, I'm like, it's it doesn't forget easily. it humans are Humans hallucinate way more than AI does.
00:27:20
danchez
It's a human problem.
00:27:21
Liza Adams
yeah
00:27:22
danchez
So I love that you're using this to like guide strategy and bring strategy back to the forefront of the tactical side of marketing.
00:27:28
Liza Adams
I don't think we have a choice.
00:27:30
danchez
Yeah, it's, I don't know, it just is somebody who i I jump, I jump between these two fields all the time between the the clouds and the dirt of marketing. So it's fun to find a process to pull it pull it together with everybody.
00:27:43
Liza Adams
And Dan, I love the way you framed it, that how many times have we seen these strategy documents be created?
00:27:43
danchez
Are
00:27:51
Liza Adams
We spend a lot of time and a lot of money doing them. And sometimes we bring in outside consultants, right? They they do tons of interviews and then they do persona development and they identify our ICP.
00:28:01
danchez
Yep.
00:28:03
Liza Adams
We we spend tens and even hundreds of thousands of dollars.
00:28:08
Liza Adams
It gets presented and then it sits. it in OneDrive or Google Drive, right?
00:28:08
danchez
Yep.
00:28:13
danchez
Yep.
00:28:15
Liza Adams
And never to be seen again. But now I have so many clients where we aren't feeding that into the work that we do consistently.
00:28:27
Liza Adams
And in the case of custom GPTs, that's the benchmark, right? Like it's in there. Without that, what does AI have to go by? Nothing. So I think it's such a good call out, you know,
00:28:41
Liza Adams
I didn't realize it until you said it. You said it. How many times have these documents just sat there being not useful?
00:28:50
danchez
Almost every day.
00:28:52
Liza Adams
Now they're front and center. Without them, it's bad marketing.
00:28:58
danchez
I'm a producer for a different AI show around sales and they they have even worse problems because they'll make quarterly objectives of things they want sales to push. But of course you want it to be pushed well two weeks ago usually, but they have issues where they just can't even, because you know, you got to, you got a whole sales force and you got to coach them, you got to train them and then you got to reinforce it. It's not enough to do a day long training with all the sales people on the new thing they they need to know.
00:29:23
danchez
you have to reinforce it and then coach it. like it's It's like six, nine, 10 months before it's actually being done, which is so slow in today's market.
00:29:29
Liza Adams
Yeah.
00:29:31
danchez
So now AI is making it possible to bring it forward for them, and that's that's something the sales people are working on, but I'm like, it's working for marketers too, because we set objectives, we set standards, and it's just not it just wasn't happening without repeated training and reinforcement, which most of us never did because there was always some fire to put out with marketing.
00:29:31
Liza Adams
Yeah.
00:29:34
Liza Adams
Yeah.
Analyzing Customer Reviews with AI
00:29:50
Liza Adams
yeah
00:29:51
danchez
but now it's actually, it's possible and it's just exciting because like strategy will actually be useful again.
00:29:59
Liza Adams
That's right.
00:30:00
danchez
yeah
00:30:00
Liza Adams
That's right.
00:30:01
danchez
tell me a little bit more, like, do you have any other case studies? These, these are two fantastic ones. What else have you done and seen so far that people should know about?
00:30:09
Liza Adams
Yeah, so let's let's just keep going down, right? So okay, we've got our strategy, we have our target segments, got our personas, we've got our campy. So let's create some assets.
00:30:21
Liza Adams
So what I love about creating assets is talking in the language that customers use and leveraging data that we have where they're actually talking to us or talking to the masses. So what I'm referring to is you know customer reviews on review sites.
00:30:45
Liza Adams
These are customers or prospects that are reviewing us and our competitors. There's a lot of good insights in there. And they're using language that they use in in those reviews and they use in their business.
00:30:55
danchez
Yeah.
00:30:59
Liza Adams
Any kind of sales calls call transcript, right? Like from Gong, wait take that.
00:31:06
danchez
Yeah.
00:31:08
Liza Adams
Any kind of customer review from creating testimonials or case studies. i you know, when we do customer success calls or customer support calls, there's a lot of insights from all those things, right?
00:31:22
danchez
yeah
00:31:28
Liza Adams
And let's just say you've got your your messaging for Persona Dan. Persona Dan has these three pillars because there are three pain points.
00:31:39
Liza Adams
And let's just take, let's make this example simple.
00:31:43
danchez
Yeah.
00:31:43
Liza Adams
Let's take some reviews from G2, review site, right?
00:31:47
danchez
Yeah.
00:31:49
Liza Adams
And we take all of our customer reviews and we say, here's a table for Persona Dan with his three pain points and our value propositions. create the messaging that we would need for Dan and give me proof points for each one of the messages based on what you can infer from the customer reviews and use exactly the language and the words that that customer that the customers are using. Do not make it up, do not paraphrase.
00:32:29
Liza Adams
fill out this table. Now there there might be 17,000 reviews, right?
00:32:36
danchez
Yeah, yeah.
00:32:37
Liza Adams
But this thing's going to comb through it. It's going to look at our pillars. It's going to try to match up. One might be around efficiency, quality, or ROI.
00:32:47
danchez
Yeah.
00:32:47
Liza Adams
then it's going to start matching it up. right Then we'll see what it comes up with. you know we We need to double check its work. And then we could say, all right, for those proof points, see if you can find qualitative.
00:33:02
Liza Adams
proof point or qualitative improvements that we can include in our messaging. So pull it out, right? It's all public information. It's G2, like anybody can go in there, you know, Captera or Trust Radius. So see what we can find and then use that. I think we get so enamored in marketing, Dan, where, you know, we we speak.
00:33:27
Liza Adams
we We like the ittys and the ables, flexibility, scalability, you know all sort highly scalable, AI driven, all sorts of things when the customers don't talk that way.
00:33:33
danchez
Yeah.
00:33:41
Liza Adams
And then when we look at the customer reviews and we just use your language, we're mirroring them. you know There's this saying that if the customer doesn't understand,
00:33:47
danchez
Yep.
00:33:51
Liza Adams
or they see language that they don't use, that the automatic impression is, oh, it must not be for me.
00:33:58
danchez
Right.
00:33:59
Liza Adams
And we don't want that, right?
00:34:01
danchez
Yeah, yeah, yeah.
00:34:01
Liza Adams
So mirroring the language is going to be super important, and there's no better place to find it than actual customer speak, actual customer reviews and interviews and transcripts.
00:34:15
Liza Adams
and And it could give you a really good head start.
00:34:18
danchez
I feel really dumb. I'm like, how have I not done this already? Because I knew this kind of stuff. I literally did a podcast just on Grok. I don't know, Grok 2 just came out. which is kind of like a gp It's almost-4 level, but it's attached to X, AKA Twitter, right?
00:34:35
danchez
And it it's I don't know, it's it's probably a combination of some old algorithm tech and AI because you can ask it like, hey, what's currently trending with insert target market? And it's ability to go and find it and then analyze it to find out what's trending and it'll bring the receipts, it'll bring the tweets that it's pulling from.
00:34:50
danchez
for you to understand the market what the market's talking about in real time is really good. But it hadn't occurred to me yet to just scrape all the reviews from something like G2 and then have a due analysis on pain points and then to bring examples with it.
00:35:07
danchez
it makes me want to like do that immediately after this interview, actually. Cause I can think of like three clients that I'm like, how have I not done that already? It's such a quick and easy thing that you can do.
00:35:17
Liza Adams
Easy.
00:35:18
danchez
You don't even have to copy
Validating AI Analyses
00:35:19
danchez
and paste it. You can literally just go to, Hey, do a search on Bing for G2 for this company and report back. Uh, cause it can do web searching through Bing and go find it.
00:35:27
Liza Adams
Yeah. Well, and and then if you just take that use case one step further into sales, because you were going down that path, right taking customer reviews, you could say, based on the customer reviews, what are the top objections?
00:35:34
danchez
Yeah. Yeah.
00:35:45
Liza Adams
about our product. Heck, that's your template for your objection handling document for sales, right?
00:35:50
danchez
Yep, yep.
00:35:53
Liza Adams
And then you feed it your value propositions, your positioning, your your messaging, your differentiation, and you say, all right, take a stab at answering those objections based on on what we've got.
00:36:05
danchez
Yep, yep.
00:36:10
Liza Adams
Again, gets you started on a draft. for crafting your objection handling FAQ for sales.
00:36:18
danchez
Yeah. Question, when you're feeding it something, like GT reviews are just long and extensive. There's usually only dozens of them per company. Sometimes it's a very popular product. You have hundreds. When it comes to companies that might have like hundreds of reviews and they're like short, maybe Amazon reviews,
00:36:35
danchez
How do you validate the AIs doing a good quantitative analysis? Because some of the problems might be coming up more frequently than others versus it like just hovering down on like maybe the first or the last one that they might give a special attention to even though it was only mentioned a few times.
00:36:51
danchez
so Do you check to see if that's consistent?
00:36:52
Liza Adams
Yeah.
00:36:54
Liza Adams
Yeah, I try not to give it too many, because at least in today's versions of the AI, and I'm not deep in the technical
00:37:01
danchez
Yeah.
00:37:04
Liza Adams
reasons why it can and can't, right? But there's also this notion of the context window and how big it is.
00:37:10
danchez
ye
00:37:11
Liza Adams
So, you know, Gemini's got like 2 million, right? So, but I've not tried to use Gemini for this. i mean I'm using chat GPT, right? So, 120K tokens.
00:37:22
Liza Adams
The other aspect of this is it's not very good in math. So once I start needing very precise information, then I really don't trust it.
00:37:34
Liza Adams
So I try, here are a couple of things. Let's just say there's 17,000 reviews. I don't give it 17,000 reviews.
00:37:42
danchez
Yeah.
00:37:42
Liza Adams
i I probably will narrow down the window to, all right, instead of the last year, just give me the last three months.
00:37:50
danchez
Okay.
00:37:51
Liza Adams
Now we don't have 17,000 reviews. Now we only have 2,000, right?
00:37:55
danchez
Yeah, you maybe take the most recent ones.
00:37:56
Liza Adams
Okay. Most recent ones.
00:37:57
danchez
Recency is always pretty big, good way to filter it.
00:37:59
Liza Adams
Yeah. Or, you know, in CAPterra and G2, you could also say, I only want the enterprise ones.
00:38:09
Liza Adams
I only want the healthcare ones.
00:38:10
Liza Adams
You know, start narrowing it down, right? Because we we do have an ICP. Because if you took everything, this thing's huge. So start, you know, really go down to the segmentation to see if you could narrow it down.
00:38:10
danchez
I see.
00:38:23
Liza Adams
And if you still can't narrow it down small enough, then what I do is like, chunking it down into pieces. Do do the do the first 500.
00:38:36
Liza Adams
I want to get insights from that. Do the second 500. Give me insights on that. And then see if you can combine them. you know Insights from the first one, insights from the second one, and then combine, if that makes sense.
00:38:44
danchez
yeah Yeah, that's that's yeah no that's how everybody gets around the context
Comparing AI Tools: ChatGPT vs Claude
00:38:51
danchez
windows. You have to do it in chunks and then summary and then combine the chunks or combine the summaries.
00:38:53
Liza Adams
In chunks.
00:38:56
Liza Adams
Yeah. And it also hi it helps for checking it too, right? Because I can't check like something super big, but I can check something that's, you know, maybe one or two pages of a spreadsheet.
00:39:00
danchez
and
00:39:09
danchez
Yeah, and that is a way to actually quantify it because if it's giving a summary of highlights of whatever you were looking for whether it's pain points or High like things that they liked about the product doesn't matter whatever you're looking for and it gives a summary And then you have it do maybe five sets of summaries Well now it can it can quantitatively look to see like well what Which which paint maybe if you went like three months and then three months back and then three months back and three months back and you could see what's trending across the year and actually quantitatively know like hey like this so this insight is coming up in each summary therefore this is a thing and it's it continues to trend
00:39:35
Liza Adams
yeahp
00:39:44
Liza Adams
Yeah. You know, the other way to check this, and in we've done this fairly consistently now, is same prompts, same data, do it with chat GPT, same prompts, same data, do it with Claude. See if there's a vast difference in what they're recommending in terms of value propositions.
00:40:08
Liza Adams
and what they're inferring from the reviews. If there's a vast difference, somebody is hallucinating somewhere, right? And that gives you a kind of a mental check. But if they're relatively similar, you know same trends, same value props, then you could feel a little bit more confident because now you've got two AIs yeah checking each other.
00:40:31
danchez
Yeah Yeah, no, that's a good way to do it I I tend not to ever use clot or any of the other ais I'm pretty committed to chat GPT just because I I find learning how to use one, it's like learning how to use them all. I just don't want to pay for them all. So I just, I just stick with chat GPT. I'm hoping eventually it'll be Microsoft co-pilot because then it'll integrate with more stuff. I don't feel like co-pilot's ready for prime time quite yet, but it's getting close.
00:40:56
Liza Adams
Well, then I will tell you that for as much as I use it, I would say that at least for my purposes, Claude is a better thought partner.
00:41:05
danchez
Yeah.
00:41:05
Liza Adams
it It's a bit more collaborative, a bit more analytical in my mind. I truly just like the way it thinks about strategy, right?
00:41:17
Liza Adams
Like if I ask it for different perspectives, it it challenges me actually, right?
00:41:23
danchez
Yeah, yeah.
00:41:24
Liza Adams
Yeah, chat GPT kind of, and it's less collaborative. it It just answers a question rather than, Claude tends to ask me questions.
00:41:36
Liza Adams
And I like that.
00:41:37
danchez
Yeah, that's interesting.
00:41:37
Liza Adams
Before I answer your question, it it asked for clarification.
00:41:38
danchez
Yeah.
00:41:41
Liza Adams
So anyway, you might want to try it.
00:41:43
danchez
Yeah. I hear, I hear, I keep hearing that Claude's like a step, a step above. I just expect that chat GPT would be launching number five or it's fifth version soon.
00:41:52
danchez
So I'm just like, just going to stick with that for now because I'm going to move over to Claude and have to move all my custom GPTs over to like the new thing.
00:41:53
Liza Adams
yeah You're just waiting. Yeah. but
Conclusion and Key Takeaways
00:42:00
danchez
And I'm like, nah, I'm just going to leave them because I have like dozens of them and I'm like, I ain't moving them. I'm not going to copy and paste them all over.
00:42:02
Liza Adams
That is the problem.
00:42:05
Liza Adams
That is the problem i that claude you can't do custom GPTs with Claude, right?
00:42:07
danchez
Yeah.
00:42:10
danchez
Can't you now? I thought they released a mechanism for that.
00:42:13
Liza Adams
Oh, did they?
00:42:13
danchez
Didn't they call it gems or some gems?
00:42:14
Liza Adams
Just recently?
00:42:15
danchez
Or maybe that was Gemini's thing. I think they do.
00:42:16
Liza Adams
a
00:42:17
danchez
I think they do have a custom GPT thing with Claude now. At least that's what I heard.
00:42:21
Liza Adams
Awesome.
00:42:22
danchez
Yeah.
00:42:22
Liza Adams
Fantastic.
00:42:23
danchez
Whether you can do all the other little tiny things. I don't know. Like, can you upload PDFs? Can you have it scrape the Reb? Can it do Excel sheets? I don't know. But I've heard it can do GPTs.
00:42:32
Liza Adams
Now, Claude, at least today, as far as I know, Claude still can't browse the web. So if it's doing a custom GPT, but it can't browse the web, we've got a little bit of a challenge.
00:42:46
danchez
Yeah. Yeah. Well, thank you so much for coming on this episode. I've learned a lot and I feel like I've been unloaded like like a full course meal, like a five course meal with like how to really unpack a whole strategy from beginning to middle to end.
00:42:57
Liza Adams
course
00:43:01
Liza Adams
and
00:43:02
danchez
And this is one of my favorite episodes so far.
00:43:05
Liza Adams
Oh, thank you.
00:43:05
danchez
I think you've really challenged me on how to use check how to use AI on a strategic level and not just the tactical level, which is what I usually tend to gravitate towards. I like the strategy, but you've really given me a framework for how to like fit it together to actually use it to really speed up and quantify and do better strategy. So thank you so much for coming on the show and sharing that.
00:43:27
Liza Adams
Yeah my pleasure this was so much fun and you've given me like little new analogies little tidbits that kind of crystallize it for people and make it more relatable and half the time it's really just that right like you know when when you say something different and then now the idea is crystallized and I'm like oh okay I didn't realize that that's what you're talking about so I really like this notion of
00:43:45
danchez
Oh, yeah.
00:43:51
Liza Adams
No strategy document will now sit and and and collect dust, right? It is now the the the the guts of what powers AI to deliver personalized and highly relevant responses and content for us.
00:44:10
danchez
So where can people to go to learn more from you to find some more about the the case studies that you've mentioned and connect with you online?
00:44:18
Liza Adams
Yeah, so I post liberally on on LinkedIn. and And the reason I do that is I'm so passionate about elevating the strategic value of marketing. And like I said, initially, I do feel that this is the gift that will turn perceptions on its head that marketing is primarily a tactical function. We are a strategic function. we Our North Star is deeply understanding the customers.
00:44:44
Liza Adams
And absent that, it's just marketing for the sake of marketing, right? So that that is my passion and that's why I'm sharing more broadly. The other thing is this AI community, I'm indebted to it. you know I learned so much from so many people. Like what you hear from me is as a result of learnings from so many, like you just gave me a bunch of key learnings today, right? And this is one of those where the AI companies are building to create God when it comes to technology, but no one is really teaching us
00:45:17
Liza Adams
the best practices because there are no best practices. It is up to us users to determine how to apply AI responsibly. So you know when when when I share, this is like part of my gratitude. you know I'm sharing because so many people have shared with me. and So LinkedIn is one. The other one is my website, growthpath.net. And you know my work is to inspire people with what's possible. And if I can help, reach out to me.
00:45:45
danchez
And of course, all the links to these will be in the description if you want to jump over to them and learn more. Thank you so much again for joining me.
00:45:53
Liza Adams
Thank you, Diana.