The AI Adoption Plateau: Why Change Management Still Rules Everything

In this episode of the FutureCraft GTM Podcast, hosts Ken Roden and Erin Mills reunite with returning favorite Liza Adams to discuss the current state of AI adoption in marketing teams. Liza shares insights on why organizations are still struggling with the same human change management challenges from a year ago, despite significant advances in AI technology. The conversation covers practical frameworks for AI implementation, the power of digital twins, and Liza's approach to building hybrid human-AI marketing teams. The episode features Liza's live demonstration in our new Gladiator segment, where she transforms a dense marketing report into an interactive Jeopardy game using Claude Artifacts.

Unpacking AI's Human Challenge

Liza returns with a reality check: while AI tools have dramatically improved, the fundamental challenge remains human adoption and change management. She reveals how one marketing team successfully built a 45-person organization with 25 humans and 20 AI teammates, starting with simple custom GPTs and evolving into sophisticated cross-functional workflows.

• The Digital Twin Strategy: Liza demonstrates how creating AI versions of yourself and key executives can improve preparation, challenge thinking, and overcome unconscious bias while providing a safe learning environment for teams.
• The 80% Rule for Practical Implementation: Why "good enough" AI outputs that achieve 80-85% accuracy can transform productivity when combined with human oversight, as demonstrated by real-world examples like translation and localization workflows.
• Prompt Strategy Over Prompt Engineering: Liza explains why following prompt frameworks isn't enough—you need strategic thinking about what questions to ask and how to challenge AI outputs for better results.

00:00 Introduction and Balance Quote 00:22 Welcome Back to FutureCraft 01:28 Introducing Liza Adams 03:58 The Unchanged AI Adoption Challenge 06:30 Building Teams of 45 (25 Humans, 20 AI) 09:06 Digital Twin Framework and Implementation 17:34 The 80% Rule and Real ROI Examples 25:31 Prompt Strategy vs Prompt Engineering 26:02 Measuring AI Impact and ROI 28:21 Handling Hallucinations and Quality Control 32:50 Gladiator Segment: Live Jeopardy Game Creation 40:00 The Future of Marketing Jobs 47:49 Why Balance Beats EQ as the Critical Skill 51:09 Rapid Fire Questions and Wrap-Up

Edited Transcript:

Introduction: The Balance Between AI and Human Skills

As AI democratizes IQ, EQ becomes increasingly important. Critical thinking and empathy are important, but I believe as marketers, balance is actually more important.

Host Updates: Leveraging AI Workflows

Ken Roden shares his approach to building better AI prompts by having full conversations with ChatGPT, exporting them to Word documents, then using that content to create more comprehensive prompts. This method resulted in more thorough market analysis with fewer edits required.

Erin Mills discusses implementing agentic workflows using n8n to connect different APIs and build systems where AI tools communicate with each other. The key insight: break workflows down into steps rather than having one agent handle multiple complex tasks.

Guest Introduction: Liza Adams on AI Adoption Challenges

Liza Adams, the AI MarketBlazer, returns to discuss the current state of AI adoption in marketing teams. Despite significant technological advances, organizations still struggle with the same human change management challenges from a year ago.

The Core Problem: Change Management Over Technology

The main issue isn't about AI tools or innovation - teams can't simply be given ChatGPT, Claude, Gemini, and Perplexity and be expected to maximize their potential. Marketing teams are being handed tools while leaders expect employees to figure out implementation themselves.

People need to see themselves in AI use cases that apply to their specific jobs. Joint learning sessions where teams share what works and what doesn't are essential. The focus has over-pivoted to "what's the right tool" when it should be on helping people understand, leverage, and make real impact with AI.

The AI Adoption Plateau

Many organizations face an AI adoption plateau where early adopters have already implemented AI, but a large group struggles with implementation. Companies attempting to "go fully agentic" or completely redo workflows in AI are taking on too much at once.

Success Story: The 45-Person Hybrid Team

Liza shares a case study of a marketing team with 45 members: 25 humans and 20 AI teammates that humans built, trained, and now manage. They started with simple custom GPTs, beginning with digital twins.

Digital Twin Strategy for AI Implementation

Digital twins are custom GPTs trained on frameworks, thinking patterns, pu

Episode Transcript

(00:00):
I've always said as AI democratizes iq, EQ becomes increasingly important, critical thinking, empathy and all those things, are important, but.

(00:09):
.508151414I believe as marketers balance is actually more important Hey crafters. 3 00:00:14,9.509151414 --> 00:00:20,759.509151414 Just a reminder, this podcast is for informational entertainment purposes only and should not be considered advice. 4 00:00:21,109.509151415 --> 00:00:30,789.510151415 The views and opinions expressed our own and do not represent those of any company or business we currently work with are associated with, or have worked with in the past. 5 00:00:31,389.510151415 --> 00:00:33,309.510151415 for tuning in to the FutureCraft podcast. 6 00:00:33,579.510151415 --> 00:00:34,549.510151415 Let's get it started. 7 00:00:34,549.510151415 --> 00:00:42,172.319675224 uh, uh, uh, um, uh, Hey there. 8 00:00:42,172.319675224 --> 00:00:50,632.319675224 Welcome to the Future Craft Go to Market podcast where we're exploring how AI changing all things go to market, from awareness to loyalty and everything in between. 9 00:00:51,142.319675224 --> 00:00:54,292.319675224 I'm Ken Roddin, one of your guides on this exciting new journey. 10 00:00:54,982.319675224 --> 00:01:00,862.319675224 And I'm Erin Mills, your co-host, and together we're gonna unpack the future of AI and go to market. 11 00:01:01,217.319675224 --> 00:01:06,467.319675224 We're gonna share some best practices, how-tos and interview industry pioneers and leaders. 12 00:01:07,307.319675224 --> 00:01:20,417.319675224 So Ken, how are you paving the way with AI and go to market this week? I don't know if I'm paving the way, but I am leveraging some things that we've learned from some of our guests so far this season. 13 00:01:21,77.319675224 --> 00:01:27,677.319675224 A couple weeks ago, Tani talked about how she approaches building a prompt and getting an answer by having a conversation with ai. 14 00:01:29,602.319675224 --> 00:01:39,542.319675224 Few weeks ago, we talked to Chris Penn and he said, the more information you could put in one single prompt or input you're going to be more successful. 15 00:01:39,592.319675224 --> 00:01:47,282.31967522 So what I did was I had a full conversation with Chacho VT about a topic copied and pasted it, exported it, put it into a word doc. 16 00:01:47,657.31967522 --> 00:01:53,867.31967522 Then used it to make a prompt and uploaded that information and I got a really solid output. 17 00:01:53,957.31967522 --> 00:02:02,747.31967522 I used it to do some market analysis on a emerging industry for a client I'm working with, and what I found was I. 18 00:02:03,77.31967522 --> 00:02:07,37.31967522 It was much more comprehensive and I had way less edits to make. 19 00:02:07,287.31967522 --> 00:02:14,217.31967522 And it even addressed some of the nuance in the conversation where it was like you're originally talking about this, but did you want it like that? And I was like, wow. 20 00:02:14,217.31967522 --> 00:02:15,297.31967522 So it really does pay attention. 21 00:02:15,457.31967522 --> 00:02:16,807.31967522 It's a big time saver. 22 00:02:16,967.31967522 --> 00:02:19,37.31967522 And also a frustration saver. 23 00:02:19,237.31967522 --> 00:02:23,607.31967522 I don't feel like I got as frustrated working on this prompt as I might have doing something this complex. 24 00:02:23,757.31967522 --> 00:02:24,387.31967522 So it was great. 25 00:02:25,17.31967522 --> 00:02:25,887.31967522 I love it. 26 00:02:26,57.31967522 --> 00:02:37,232.31967522 How about you? I did something similar in terms of taking one of the plays outta the playbooks that our guest offered, who was Chase Hannigan a couple weeks back. 27 00:02:37,312.31967522 --> 00:02:44,912.31967522 And being able to connect Tali with some of the other APIs and build these, essentially. 28 00:02:45,517.31967522 --> 00:02:51,457.31967522 Entire systems that continue to talk to each other and have different roles is something I've been really interested in working on. 29 00:02:51,507.31967522 --> 00:02:58,767.31967522 I had already created the assistant that he had been talking about in the episode, but I added a lot of bells and whistles to what it can do. 30 00:02:59,7.31967522 --> 00:03:06,147.31967522 And so now it does a lot more searching for me and it scrapes different websites that I'm interested in. 31 00:03:06,177.31967522 --> 00:03:08,667.31967522 And I've been really happy with the results so far. 32 00:03:09,547.31967522 --> 00:03:15,137.31967522 Do you feel like you're getting better with N eight N now that you've been using it? I do. 33 00:03:15,137.31967522 --> 00:03:22,397.31967522 I think more than getting better with N eight N is understanding how the whole ENT systems work together. 34 00:03:22,667.31967522 --> 00:03:33,947.31967522 And so when I go over to relevance or I'm using another tool or building on lovable and using Super base, I feel like the NANN eight N helps to really. 35 00:03:34,922.31967522 --> 00:03:46,262.31967522 It helps you to visualize how all these things work together so that when you're working in these different systems, it actually ties the story together, which might make sense or might not, but in my head it makes sense. 36 00:03:46,262.31967522 --> 00:04:03,662.31967522 it's a tip that could really help people who are moving from using, chat BT and thinking about moving to more of an agentic workflow you gotta break it down into steps, right? You don't wanna have one agent doing your whole social media strategy, your content calendar, and doing the LinkedIn posts, that's just too much for it to do. 37 00:04:03,662.31967522 --> 00:04:07,752.31967522 So you've gotta break it down into different sections or different nodes, wow. 38 00:04:09,492.31967522 --> 00:04:10,692.31967522 I think that's exactly it, Ken. 39 00:04:10,697.31967522 --> 00:04:20,866.31967522 And especially someone like me I'm a super visual learner and so just being able to see how things come together helps me to really grasp the concept. 40 00:04:21,76.31967522 --> 00:04:30,496.31967522 Even with relevance, even though it's easier, I think for most probably like an entry point for people to use when things go wrong, I have a harder time navigating how to fix it. 41 00:04:30,546.31967522 --> 00:04:43,676.31967522 With NAN and some of these other tools where you can really visually understand what's connecting with the, or what's connecting with each other, it can give you a lot more comfort or give you a lot more of an understanding of how things are going. 42 00:04:44,201.31967522 --> 00:04:56,51.31967522 Hey, who are we talking to today, by the way? Oh, we're talking to one of our favorites from last season, Liza Adams, who we all know has great frameworks and is a prolific, a LinkedIn poster. 43 00:04:56,411.31967522 --> 00:04:58,241.31967522 I can't wait to talk to her. 44 00:04:58,241.31967522 --> 00:05:00,961.31967522 As she's a good friend of the pod, we'll be chatting with her coming up. 45 00:05:01,531.31967522 --> 00:05:03,661.31967522 I'm so excited. 46 00:05:03,661.31967522 --> 00:05:07,501.31967522 I've been waiting for this for so long, so actually I can't wait it anymore. 47 00:05:07,501.31967522 --> 00:05:17,517.00538951 We've gotta go over and talk to her uh, uh, uh, Today we're reuniting with one of our very favorites, Liza Adams. 48 00:05:17,517.00538951 --> 00:05:23,457.00538951 Welcome back the AI MarketBlazer who showed us how pilot wins can transform go to market. 49 00:05:23,727.00538951 --> 00:05:25,887.00538951 Liza, really excited to hear what you have for us today. 50 00:05:26,817.00538951 --> 00:05:28,407.00538951 I couldn't wait to come back. 51 00:05:29,127.00538951 --> 00:05:35,547.00538951 I was waiting for Erin and Ken to call me and say, Hey, would you come back? So super thrilled to be here. 52 00:05:36,162.00538951 --> 00:05:37,152.00538951 Thanks for joining us. 53 00:05:38,202.00538951 --> 00:06:08,282.00538951 So it's been about a year since we last spoke with you and since the sort of AI Marketing Masterclass episode aired, what surprised you the most on how AI really hit or missed for marketers in 2024 and leading into 2025? You know what surprised me is the fact that we are still having very similar conversations as a year ago when it comes to. 54 00:06:09,782.00538951 --> 00:06:11,102.00538951 Human change management. 55 00:06:12,102.00538951 --> 00:06:27,722.00538951 I think we tried to land that last year that this isn't about tools or AI innovation or technology, that we can't just simply give people Chachi, PT, clot and Gemini and perplexity and expect them to get the most out of it. 56 00:06:27,812.00538951 --> 00:06:32,12.00538951 I am still seeing marketing teams getting handed. 57 00:06:32,37.00538951 --> 00:06:38,167.00538951 Tools and the leaders are expecting the teams to figure it out themselves. 58 00:06:39,387.00538951 --> 00:06:42,247.00538951 People have to see themselves in it. 59 00:06:42,992.00538951 --> 00:06:45,542.00538951 use cases that apply to their jobs. 60 00:06:46,32.00538951 --> 00:06:51,712.00538951 Have joint learning sessions, sharing what works and what doesn't work. 61 00:06:52,112.00538951 --> 00:06:56,102.00538951 I think we have over pivoted on what's the right tool. 62 00:06:56,102.00538951 --> 00:07:05,602.00538951 There's this and that tool and the other, right? When the pivot should be more on how do we help people better understand leverage and truly make an impact with ai. 63 00:07:08,62.00538951 --> 00:07:32,472.00538951 Yeah, I really find that we are still in this same conversation that we were having a year ago in many areas something I'm calling the AI adoption Plateau, where the people who are gonna adopt AI have, and we still now have this bucket of a big group of people saying, we are trying this stuff, but why isn't it working? You hit on the exact point we've got this change management issue, and one of the reasons I am seeing. 64 00:07:33,237.00538951 --> 00:07:40,167.00538951 People struggle with it is their company was like, let's go fully agentic, or let's just redo all of our workflows in ai. 65 00:07:40,167.00538951 --> 00:07:40,977.00538951 And that's too much. 66 00:07:41,257.00538951 --> 00:07:45,207.00538951 It reminded me that you recommend founders start with Tiny AI win. 67 00:07:45,237.00538951 --> 00:08:13,297.00538951 Could you share a success story that really illustrates that approach? What worked? And maybe highlight some pitfalls for people to avoid? Yeah, so I have, been talking about one of the key use cases of a success were we now have a marketing team that is 45 members strong, and 25 of them are humans, and then 20 of 'em are AI teammates that the human beings actually built, trained, and are now managing and maintaining. 68 00:08:13,367.00538951 --> 00:08:14,807.00538951 We didn't get there overnight. 69 00:08:16,577.00538951 --> 00:08:20,87.00538951 One of, one of the key successes was we started with. 70 00:08:20,417.00538951 --> 00:08:28,727.00538951 Super simple custom gpt, and one of those super simple custom gpt was a digital twin. 71 00:08:28,727.00538951 --> 00:08:34,17.00538951 So you know, Liza, A GPT as an example, right? In dice. 72 00:08:34,17.00538951 --> 00:08:37,137.00538951 That's the company that, that, that's transformation. 73 00:08:37,497.00538951 --> 00:08:46,917.00538951 The CMO the head of marketing, the head of demand generation built digital twins, not just of themselves, but of the executive team. 74 00:08:48,117.00538951 --> 00:08:49,137.00538951 So basically. 75 00:08:49,497.00538951 --> 00:09:03,97.00538951 Trained custom GPTs on their frameworks, their thinking their anything that's publicly available that they've written their Myers Briggs, all sorts of things. 76 00:09:03,97.00538951 --> 00:09:14,502.00538951 So this GPT basically knows them super well, right? The reason why I'm calling it simple is we know ourselves best than better than anybody else, right? We could check its work. 77 00:09:14,562.00538951 --> 00:09:17,232.00538951 We know when it doesn't sound like us. 78 00:09:17,472.00538951 --> 00:09:20,982.00538951 We know when It doesn't sound like a strategy we've done. 79 00:09:21,492.00538951 --> 00:09:26,802.00538951 And what I love about it is it's designed not to mimic us. 80 00:09:27,192.00538951 --> 00:09:32,112.00538951 It's designed to learn about us so that it can find our blind spots. 81 00:09:32,812.00538951 --> 00:09:36,412.00538951 So that it can challenge our thinking because we get into our rut. 82 00:09:36,412.00538951 --> 00:09:37,462.00538951 We always think this way. 83 00:09:37,462.00538951 --> 00:09:40,972.00538951 We also have unconscious bias, all sorts of things. 84 00:09:40,972.00538951 --> 00:09:47,782.00538951 So the GPTs were designed to make ourselves better, right? To actually compliment us and overcome our weaknesses. 85 00:09:48,352.00538951 --> 00:09:53,812.00538951 On the other side, where we're replicating somebody else, like the executive de. 86 00:09:54,792.00538951 --> 00:09:59,892.00538951 We use that to prepare for any kind of conversation that we're going to have with them. 87 00:09:59,952.00538951 --> 00:10:08,82.00538951 So Megan, in her case, when she has to present a campaign to the CMO and the CFO, she says, here's my campaign deck. 88 00:10:08,617.00538951 --> 00:10:14,127.00538951 Help me understand what questions they might ask, fill in some gaps. 89 00:10:14,127.00538951 --> 00:10:26,217.00538951 What did I not address? So she either fills in those gaps based on what the GPT says, or mentally she's thinking, all right, I know that they're gonna ask me these questions based on what the GPT said. 90 00:10:26,367.00538951 --> 00:10:29,337.00538951 I'm going to prepare a response to that question. 91 00:10:29,707.00538951 --> 00:10:32,217.00538951 I think it's just such a brilliant starting point. 92 00:10:32,517.00538951 --> 00:10:36,597.00538951 And from there you could see how it got. 93 00:10:37,137.00538951 --> 00:10:41,157.00538951 More complex, more more tied to workflows. 94 00:10:41,157.00538951 --> 00:10:53,157.00538951 They begin using GPTs, not just to do very specific tasks, like they have a pitch deck builder, they have a content topic, ideator, they have a campaign performance analyzer. 95 00:10:53,587.00538951 --> 00:11:08,297.00538951 Now Megan, who's the head of demand Generation, she has actually strung these gpt and chained them together so that it works not just within marketing, but also begins to work with sales and cs. 96 00:11:08,687.00538951 --> 00:11:14,27.00538951 So if you can imagine you got a campaign launch, GPT chain to a. 97 00:11:15,292.00538951 --> 00:11:24,472.00538951 Sales enabler, GPT chained into a question and answer GPT to help CES answer customer questions. 98 00:11:24,832.00538951 --> 00:11:32,932.00538951 Now, she's not no longer just working in marketing, she's working on workflow that cuts across multiple organizations. 99 00:11:33,142.00538951 --> 00:11:36,182.00538951 Stop anyway from a simple, digital twin. 100 00:11:36,602.00538951 --> 00:11:44,722.00538951 To something more complex by function, to know something that chains across multiple organizations in one workflow. 101 00:11:46,352.00538951 --> 00:11:52,982.00538951 I think that's such a great example and something that people can get started with today without having to have a lot of technical expertise. 102 00:11:53,532.00538951 --> 00:12:22,482.00538951 If you think about when you even work with a lot of clients, talking to a lot of folks thinking about aI magnifying some of the challenges, like identifying some of those gaps? 'cause you talked about it, uploading a PowerPoint, but how have you seen AI amplify some of those, whether they're pockets in go to market strategy how can teams avoid the trap of making existing challenges worse? Yeah, so the, there's many examples and I think there's this. 103 00:12:23,217.00538951 --> 00:12:31,837.00538951 Preconceived notion that when we follow a prompt framework the answer is going to be really good. 104 00:12:32,527.00538951 --> 00:12:32,737.00538951 Yeah. 105 00:12:32,767.00538951 --> 00:12:39,47.00538951 But, and prompting, like they say prompt engineering, you need to be good at prompting and all sorts of things. 106 00:12:39,617.00538951 --> 00:12:47,117.00538951 But AI is a magnifier, to your point, Erin, right? It's a magnifier of anything good, bad, or indifferent. 107 00:12:47,477.00538951 --> 00:12:49,127.00538951 And I'll come back to the prompting piece. 108 00:12:50,267.00538951 --> 00:12:54,137.00538951 So let's just say that you follow a prompt framework. 109 00:12:54,137.00538951 --> 00:12:56,747.00538951 I don't care if it's open AI or Gemini. 110 00:12:57,267.00538951 --> 00:12:59,337.00538951 Microsoft is a prompting framework. 111 00:12:59,437.00538951 --> 00:13:25,497.00538951 Claude Christopher Penn has a prompting framework called race, and then I've actually amended his prompting framework and turned it into grace because the G stands for goals, R for role, A for action, C for context, and then E for examples, if you prompt the AI enough with that framework and you give it all that information, then your expectation is that it's gonna give you good results. 112 00:13:26,127.00538951 --> 00:13:27,687.00538951 Here's what's happening. 113 00:13:27,717.00538951 --> 00:13:31,17.00538951 People follow the prompt framework, doesn't matter what it is. 114 00:13:31,17.00538951 --> 00:13:43,227.00538951 And there are lots of good ones out there, and there's still disappointed with the results, right? And they're disappointed because it's magnifying what's there and what's there is something very basic. 115 00:13:43,717.00538951 --> 00:13:45,247.00538951 It's basic thinking. 116 00:13:45,517.00538951 --> 00:13:48,727.00538951 It follows the framework, but the thinking is still basic. 117 00:13:49,57.00538951 --> 00:13:52,637.00538951 So for example, let's just say you're having a churn problem. 118 00:13:52,677.00538951 --> 00:13:55,107.00538951 You have goals and all sorts of things. 119 00:13:55,107.00538951 --> 00:13:59,117.00538951 You give it context that this is for, travel management software. 120 00:13:59,117.00538951 --> 00:14:01,847.00538951 We're experiencing this churn and here's some data. 121 00:14:02,257.00538951 --> 00:14:05,227.00538951 And here's some examples of what we've experienced. 122 00:14:05,617.00538951 --> 00:14:09,367.00538951 Help me figure out how to reduce churn, okay. 123 00:14:09,772.00538951 --> 00:14:10,642.00538951 Awesome, prompt. 124 00:14:10,852.00538951 --> 00:14:11,332.00538951 Great. 125 00:14:12,292.00538951 --> 00:14:18,142.00538951 However, we are simply asking it a very basic question of how to reduce churn. 126 00:14:18,142.00538951 --> 00:14:50,682.00538951 We're not forcing it to think, right? If we said things like, Hey, can you challenge me and think about this differently? 'cause I'm assuming that it's a churn problem, but could it perhaps be a different kind of problem? Can you look at this data? And determine whether it's a churn problem or are there potentially other opportunities that this is telling us? Because an example that I recently posted about this, a churn problem could actually be an opportunity for an upsell. 127 00:14:52,57.00538951 --> 00:14:55,862.00538951 A hundred percent, right? Like we're seeing a lot of churn. 128 00:14:55,862.00538951 --> 00:15:03,492.00538951 Customers are leaving why are they leaving? A traditional response to that question is, Hey, you need a loyalty program. 129 00:15:03,492.00538951 --> 00:15:06,762.00538951 You need, better Cs, or all sorts of things. 130 00:15:07,122.00538951 --> 00:15:12,192.00538951 But if you actually see it as an indicator that it's an upsell opportunity. 131 00:15:12,552.00538951 --> 00:15:18,882.00538951 Then it's a very different motion or a different response from the AI that says, oh, it's an upsell opportunity. 132 00:15:18,882.00538951 --> 00:15:32,832.00538951 Let's think about potential product offerings or potential packaging that would include new features, new functionality, or we probably need a campaign and let them know that it's coming in the next three months or something like that. 133 00:15:33,192.00538951 --> 00:15:37,422.00538951 And now it's no longer a churn problem, it is now an upsell. 134 00:15:37,637.00538951 --> 00:15:38,537.00538951 Opportunity. 135 00:15:38,747.00538951 --> 00:15:41,497.00538951 So it's not necessarily prompt engineering. 136 00:15:41,617.00538951 --> 00:15:42,937.00538951 It's prompt strategy. 137 00:15:43,417.00538951 --> 00:15:43,687.00538951 Yeah. 138 00:15:44,27.00538951 --> 00:16:04,842.00538951 It is what is the strategy? What is the thinking we want to elevate in this conversation? I totally agree with that the critical thinking piece is something we've been talking a lot about you think about kids growing up now and what do they need to learn? It's not necessarily just how to prompt, but that critical thinking, what's the problem? That diagnosis is something. 139 00:16:05,212.00538951 --> 00:16:11,692.00538951 One of the highlights last year we talked about, which I think ties well, is the, aspire, align and implement. 140 00:16:12,152.00538951 --> 00:16:12,932.00538951 I'm curious. 141 00:16:12,942.00538951 --> 00:16:32,27.00538951 Having implemented that with more teams over the last year, what has been the most critical for teams and why? Yeah, and I touched on this earlier we can't underestimate the amount of work required to meet people where they're at and bring them along. 142 00:16:33,47.00538951 --> 00:16:34,367.00538951 Cannot underestimate it. 143 00:16:34,467.00538951 --> 00:16:38,347.00538951 We can have the most awesome MarTech stack out there. 144 00:16:39,307.00538951 --> 00:16:41,497.0053895 We can have four ais. 145 00:16:41,677.0053895 --> 00:16:53,492.0053895 All sorts of support from vendors, but when we don't understand the fears, the motivations, the challenges of people, and we don't give them the space to learn. 146 00:16:54,622.0053895 --> 00:17:00,682.0053895 We're all pretty oversubscribed, right? Like anywhere from 120 to 150%, maybe beyond. 147 00:17:00,772.0053895 --> 00:17:03,172.0053895 There's never enough hours in the day. 148 00:17:03,652.0053895 --> 00:17:10,62.0053895 And then we're expected to learn AI on our own and no one's teaching us, oh my gosh. 149 00:17:10,62.0053895 --> 00:17:15,42.0053895 The pressure of that is immense, right? Especially when there are people saying. 150 00:17:15,582.0053895 --> 00:17:18,462.0053895 Hey, AI is gonna take your job you're way behind. 151 00:17:18,462.0053895 --> 00:17:29,2.0053895 I also don't like that, it's fear-based, right? Not only are we now oversubscribed and under pressure, we now get paralyzed by fear. 152 00:17:29,542.0053895 --> 00:17:38,932.0053895 So this whole notion of if we want to drive change, we can't underestimate the amount of work that it would require to bring people along. 153 00:17:39,622.0053895 --> 00:17:48,792.0053895 And you know how those little cars, the little matchbox cars that you pull back and then when you pull back it gains momentum and energy, and then when you let go, it goes. 154 00:17:49,242.0053895 --> 00:17:56,952.0053895 That's like analogy I'm putting in my head, let's give people the space to pull back, give them the space to learn and understand. 155 00:17:56,952.0053895 --> 00:17:57,492.0053895 Use ai. 156 00:17:58,617.0053895 --> 00:18:03,897.0053895 And then when they're inspired and we show them what's possible, let it go and then it will move forward. 157 00:18:03,947.0053895 --> 00:18:13,698.0053895 And I think right now what's happening is we're pushing the car along very manually when we can just take a little bit of time to pull back and let it go. 158 00:18:16,52.0053895 --> 00:18:30,792.0053895 Yeah, that analogy really stands out to me as something that deals with the relationship between employees and their manager, right? The how much their manager's pushing them based off whatever pressure they're getting and whatever the leadership is experiencing. 159 00:18:30,792.0053895 --> 00:18:44,22.0053895 And I recently did some research and found talking to white collar professionals that something like 93% of white collar professionals were willing to adopt AI at work, but only 50% of them trusted their. 160 00:18:44,407.0053895 --> 00:18:55,527.0053895 The leader's decision making and having an AI strategy, one of the obstacles is the change management around, I need time to learn this thing, but it's also I don't really know how you're going to use it or if you know what you're doing. 161 00:18:55,717.0053895 --> 00:18:59,617.0053895 it circles around a lot of, bit of the dynamics around change management. 162 00:19:00,337.0053895 --> 00:19:35,612.0053895 Another thing that people look at is this idea of how do we measure this as working? When you've been working with marketing teams and getting them on the path to implementing AI and integrating it into their workflows, is there a metric that you can use as a leading indicator? And then are there some metrics longer term that people can look for when they're starting to look for the impact and the outcomes? I get this question a lot because people are always saying, measure what's the ROI and all sorts of things, right? We can, get super complex with that and our heads will explode when we start thinking about all the different possibilities. 163 00:19:36,302.0053895 --> 00:19:50,127.0053895 I always guide marketing leaders to do one of three things, or in combination, right? One is to align jobs to be done to specific strategic initiatives that the company already has. 164 00:19:50,907.0053895 --> 00:20:03,267.0053895 Those jobs to be done could be a combination of use cases using ai, right? When we align it to strategic initiatives, we have a better shot at measurements because those strategic initiatives have goals. 165 00:20:03,267.0053895 --> 00:20:04,497.0053895 They have KPIs. 166 00:20:04,907.0053895 --> 00:20:12,997.0053895 And we have a better shot at adoption because it has budget, it has resources, it has the eyes of the executive team, right? You have no choice. 167 00:20:13,807.0053895 --> 00:20:15,487.0053895 You align to the jobs to be done. 168 00:20:15,487.0053895 --> 00:20:17,347.0053895 That's aligned to the strategic initiative. 169 00:20:17,347.0053895 --> 00:20:23,377.0053895 You're gonna make this thing work and you're gonna be accountable for measuring and reporting on impact. 170 00:20:23,977.0053895 --> 00:20:30,857.0053895 So that's one of the ways to force the issue, right? The second thing that I guide marketing leaders towards is. 171 00:20:31,962.0053895 --> 00:20:41,382.0053895 Pick the area of biggest pain, right? Where you have the biggest pain, you will put the resource towards it. 172 00:20:41,672.0053895 --> 00:20:56,662.0053895 You will use AI to help overcome that pain, and your goal is to alleviate the pain, right? Whether that pain is you're spending too much money with an agency or the pain is, my team is working. 173 00:20:56,952.0053895 --> 00:20:59,82.0053895 80 hours a week each person. 174 00:20:59,82.0053895 --> 00:21:00,282.0053895 Whatever it might be. 175 00:21:00,462.0053895 --> 00:21:06,402.0053895 Or the quality is so bad because we're doing this, not in a good way, There, there's not a process for it. 176 00:21:06,922.0053895 --> 00:21:08,812.0053895 I've used this example a number of times. 177 00:21:08,862.0053895 --> 00:21:11,252.0053895 One of the CMOs I worked with her and her team's. 178 00:21:11,252.0053895 --> 00:21:18,452.0053895 Biggest pain was translation and localization because they were having to do it for eight different languages. 179 00:21:19,907.0053895 --> 00:21:22,157.0053895 Every single customer facing document. 180 00:21:22,857.0053895 --> 00:21:29,857.0053895 Tens of thousands of dollars in agency fees to translate and localize, these documents. 181 00:21:30,247.0053895 --> 00:21:33,627.0053895 So guess what the team did, custom GPTs, eight of 'em. 182 00:21:34,407.0053895 --> 00:21:35,697.0053895 One per language. 183 00:21:35,697.0053895 --> 00:21:38,287.0053895 Field marketers that are native speakers. 184 00:21:38,827.0053895 --> 00:21:41,937.0053895 It doesn't mean that the custom GPTs were, perfect. 185 00:21:41,937.0053895 --> 00:21:44,877.0053895 Like it put a document on outcomes like a German. 186 00:21:45,152.0053895 --> 00:21:47,42.0053895 Translate a document all perfect. 187 00:21:47,282.0053895 --> 00:22:12,272.0053895 No, but it's like anywhere from 80 to 95% there to get that extra 5% Just human oversight versus the weeks of time and the dollars that you would need to invest with an outside agency that's super quick, right? And in that company and in that marketing team, when they said over two weeks when we built this custom gpt, we saved tens of thousands of dollars a month. 188 00:22:12,512.0053895 --> 00:22:25,642.0053895 They're like, heck, if we could do that with localization and translation, let's think about all the other gpt, because that's already like it, it was such a big ROI to them, 20 bucks a month, three people doing it. 189 00:22:26,632.0053895 --> 00:22:29,272.0053895 And then you're saving tens of thousands of dollars a month. 190 00:22:29,572.0053895 --> 00:22:34,672.0053895 You can now begin to think about all of these other use cases where you have areas of biggest pain. 191 00:22:34,822.0053895 --> 00:22:39,802.0053895 And then the last one is really around leaning into your trailblazers. 192 00:22:40,162.0053895 --> 00:22:48,967.0053895 Those who are intensely curious because they're already building the AI tools, the ai, the GPT, the gems or whatnot. 193 00:22:49,432.0053895 --> 00:22:55,782.0053895 To help themselves, with efficiency, with effectiveness, and starting to reimagine the work. 194 00:22:56,382.0053895 --> 00:23:02,142.0053895 See how they're performing already, because I bet you that they're already achieving some of these benefits. 195 00:23:02,322.0053895 --> 00:23:10,207.0053895 So those are just some like really simple ways to get started and get some numbers on the board and say, here's what we're seeing with ai. 196 00:23:11,392.0053895 --> 00:23:17,142.0053895 I think that's such a good point, Liza the folks that are the trailblazers, I think for a while were just a little bit like. 197 00:23:17,997.0053895 --> 00:23:25,857.0053895 Flying under the radar, right? How can I maximize my output without maximizing the effort I have to put in? I know a couple of those. 198 00:23:25,947.0053895 --> 00:23:33,797.0053895 One thing that people keep talking about is some early struggles with ai, which is around hallucinations and quality control. 199 00:23:34,47.0053895 --> 00:23:42,777.0053895 What practical steps do you have to ensure that AI generated content aligns with those brand standards, and how can you limit or. 200 00:23:42,787.0053895 --> 00:23:45,577.0053895 Protect yourself from horrible hallucinations. 201 00:23:46,337.0053895 --> 00:23:49,567.0053895 What was that report? Maybe Erin, you can remember. 202 00:23:49,777.0053895 --> 00:23:56,277.0053895 There's some report from OpenAI I believe, that says that AI models hallucinate anywhere from 30 to 80% of the time. 203 00:23:56,757.0053895 --> 00:23:57,897.0053895 And I was like, holy cow. 204 00:23:57,897.0053895 --> 00:23:58,407.0053895 80%. 205 00:23:59,337.0053895 --> 00:24:01,137.0053895 Yeah, that's a lot. 206 00:24:01,137.0053895 --> 00:24:01,587.0053895 That's right. 207 00:24:01,827.0053895 --> 00:24:02,517.0053895 But here's. 208 00:24:03,827.0053895 --> 00:24:30,377.0053895 You'll observe that kind of hallucination if you are using it like a question and answer machine or like a Jeopardy partner because it's not like a search engine where it's got a database and you just go into it, right? It's not a rag model, right? So if you're asking it, what is the population of Nigeria, when did Elvis die? All sorts of things like trivial questions. 209 00:24:31,97.0053895 --> 00:24:47,207.0053895 The probability of a hallucination is higher, but if you have more nuanced questions, your strategic questions, what if Scenario analysis, that's what it's designed to do, right? It's designed to brainstorm with you and improve your thinking. 210 00:24:47,507.0053895 --> 00:24:52,102.0053895 It's not designed to have exacting answers for very specific questions. 211 00:24:52,722.0053895 --> 00:24:57,997.0053895 So if you're seeing a lot of hallucinations, then just start thinking about what you're using it for. 212 00:24:58,672.0053895 --> 00:25:02,302.0053895 Because more than likely, you're probably using it as a question and answering machine. 213 00:25:02,992.0053895 --> 00:25:14,92.0053895 But to your point, Ken, there are other areas where hallucinations do occur, right? One of the things I try to do to prevent it, and there's several things. 214 00:25:14,92.0053895 --> 00:25:18,832.0053895 So the first one is check your conversations and your data set. 215 00:25:19,252.0053895 --> 00:25:36,932.0053895 If it's getting super big, like super huge data sets or very long conversations, what happens is we run into the context window limitation, which is just fancy way of saying it's the amount of information AI can remember in any give or conversation. 216 00:25:37,172.0053895 --> 00:25:40,142.0053895 So the longer the conversation, the more data it has to ingest. 217 00:25:40,502.0053895 --> 00:25:44,942.0053895 Pretty soon as you reach that threshold of what they can remember. 218 00:25:45,407.0053895 --> 00:25:56,987.0053895 It starts making up stuff confidently, to the extent that we can limit the dataset, limit your conversations, make them shorter, and then, the probability of hallucinations. 219 00:25:57,57.0053895 --> 00:26:01,667.0053895 Go down the second thing is use multiple ais. 220 00:26:03,42.0053895 --> 00:26:06,672.0053895 I think I've said this before, I have a conversation with Chad, GPT. 221 00:26:07,12.0053895 --> 00:26:29,52.0053895 cut and paste that conversation, put it in the Claude, ask Claude what it thinks, cut and paste, put it into Gemini, and they essentially check each other, right? And then I can God forbid that all three of 'em hallucinate at the same time, but when I've got three of them that I can, I have a higher probability of knowing who's hallucinating, right? So that's the second thing. 222 00:26:29,742.0053895 --> 00:26:35,622.0053895 Then one of the things that I have done now very consistently is what? It gives me an answer. 223 00:26:35,682.0053895 --> 00:26:40,892.0053895 Let's say, it's doing competitive analysis, and I say, give me your top three insights. 224 00:26:41,822.0053895 --> 00:26:42,722.0053895 I don't stop there. 225 00:26:42,992.0053895 --> 00:26:49,322.0053895 I say, for each of these insights, tell me your level of confidence, low, medium, high. 226 00:26:50,232.0053895 --> 00:26:55,902.0053895 Tell me what assumptions you made and give me your rationale for why you rated. 227 00:26:56,592.0053895 --> 00:27:08,362.0053895 Your confidence level the way you did, for the insights where you have low to medium confidence, tell me what other information might you need to increase your confidence. 228 00:27:09,562.0053895 --> 00:27:11,752.0053895 That's the prompt, right? Oh, I like that. 229 00:27:11,932.0053895 --> 00:27:11,992.0053895 Yeah. 230 00:27:13,402.0053895 --> 00:27:14,782.0053895 So I just posted that today. 231 00:27:14,907.0053895 --> 00:27:16,617.0053895 And we could put that in the show notes. 232 00:27:17,787.0053895 --> 00:27:21,357.0053895 You would be amazed how many times, let's say there's five insights. 233 00:27:22,152.0053895 --> 00:27:32,952.0053895 There's some low and medium confidence responses and it says, I have low confidence in this because I only evaluated your data. 234 00:27:33,52.0053895 --> 00:27:34,762.0053895 I did not do any web search. 235 00:27:34,762.0053895 --> 00:27:36,682.0053895 We've only looked at three competitors. 236 00:27:36,862.0053895 --> 00:27:39,622.0053895 Like it tells you, I would be more confident if. 237 00:27:40,447.0053895 --> 00:27:54,597.0053895 We had some case studies, all sorts of things, right? And I'm like, dang, this is exactly what I needed because it helps me as a human being validate, right? On the ones that are low to medium confidence versus trying to figure out everything. 238 00:27:55,47.0053895 --> 00:27:56,607.0053895 All right, something new for this year. 239 00:27:56,637.0053895 --> 00:27:59,847.0053895 Liza is the gladiator segment, which I'm heard Sure. 240 00:27:59,847.0053895 --> 00:28:01,707.0053895 You've heard on other episodes. 241 00:28:01,707.0053895 --> 00:28:03,327.0053895 And so now you are up. 242 00:28:03,607.0053895 --> 00:28:09,467.0053895 You actually turned me on to Claude Artifacts, love them and remixing pieces to make something new. 243 00:28:09,617.0053895 --> 00:28:11,57.0053895 So I'm gonna put you on the spot. 244 00:28:11,777.0053895 --> 00:28:16,707.0053895 Can you show our audience how you do it live? Yeah, I will try. 245 00:28:16,807.0053895 --> 00:28:19,177.0053895 And this one's gonna be a fun one. 246 00:28:20,197.0053895 --> 00:28:22,657.0053895 And I will share my screen. 247 00:28:24,127.0053895 --> 00:28:29,167.0053895 You guys can't make fun of my tabs, okay? Because I have way too many tabs. 248 00:28:29,827.0053895 --> 00:28:32,257.0053895 So do you see my screen? Yeah. 249 00:28:32,317.0053895 --> 00:28:37,122.0053895 So it's not gonna be exactly live, but you'll see some fun things in here. 250 00:28:37,752.0053895 --> 00:28:45,662.0053895 There is this brand new report from the Marketing AI Institute called the 2025 State of Marketing AI report. 251 00:28:45,662.0053895 --> 00:28:47,882.0053895 It's got some really good insights. 252 00:28:47,932.0053895 --> 00:28:53,972.0053895 They surveyed a number of marketers across industries, across different company sizes. 253 00:28:54,332.0053895 --> 00:28:55,712.0053895 It's a very long report. 254 00:28:56,762.0053895 --> 00:29:07,167.0053895 So what I did was normally I just put it into notebook, lm, and I have notebook lm, and I have it in my ear and it just basically tells me about key insights and things like that. 255 00:29:07,467.0053895 --> 00:29:12,367.0053895 But I wanted to do something more fun with the state of marketing AI report. 256 00:29:12,367.0053895 --> 00:29:20,747.0053895 And, I was thinking about a use case when we have offsites and we wanna do a really cool icebreaker, but at the same time. 257 00:29:20,847.0053895 --> 00:29:26,927.0053895 Make it educational, or we wanna make sure that we have interaction between the team. 258 00:29:26,927.0053895 --> 00:29:30,97.0053895 So I wanted to see what Claude can help me do. 259 00:29:31,207.0053895 --> 00:29:36,247.0053895 To turn this into something interactive that can be used by the team and make it educational. 260 00:29:36,597.0053895 --> 00:29:42,567.0053895 What you see here is the state of marketing AI report, PDFI simply uploaded that into Claude. 261 00:29:42,567.0053895 --> 00:29:44,617.0053895 This is Claude sonnet four. 262 00:29:45,157.0053895 --> 00:29:46,297.0053895 And here's my prompt. 263 00:29:46,607.0053895 --> 00:29:52,907.0053895 I say, please turn this report into a highly engaging and visually appealing interactive jeopardy game. 264 00:29:53,737.0053895 --> 00:29:55,987.0053895 Four colors blue, orange, gray and white. 265 00:29:55,987.0053895 --> 00:29:58,477.0053895 That just happens to be my brand colors. 266 00:29:58,817.0053895 --> 00:30:01,127.0053895 Then I said, please output the app itself. 267 00:30:01,247.0053895 --> 00:30:13,47.0053895 You can see that it gives, some messaging around what it's doing in terms of the game features and the game structure the different categories in the Jeopardy game and so on and so forth. 268 00:30:13,137.0053895 --> 00:30:14,637.0053895 Scoring system. 269 00:30:15,507.0053895 --> 00:30:20,487.0053895 Some data points covered, and these are from the Marketing AI Institute, but. 270 00:30:21,712.0053895 --> 00:30:23,67.0053895 I will show you what it output. 271 00:30:23,277.0053895 --> 00:30:25,737.0053895 So here's the Jeopardy game. 272 00:30:26,457.0053895 --> 00:30:27,597.0053895 So fun. 273 00:30:28,377.0053895 --> 00:30:32,367.0053895 Does this, so your conversation is still on the left hand side. 274 00:30:32,367.0053895 --> 00:30:34,287.0053895 The Jeopardy game is on the right hand side. 275 00:30:34,327.0053895 --> 00:30:41,887.0053895 It doesn't just output this, what you actually see is while I'm waiting for the Jeopardy game it codes. 276 00:30:41,987.0053895 --> 00:30:44,87.0053895 It codes all of these things. 277 00:30:44,377.0053895 --> 00:30:47,127.0053895 I didn't code this, I'm not there sitting, typing this thing. 278 00:30:47,157.0053895 --> 00:30:56,772.0053895 Oh, I typed this thing right here on the left side the prompt that I just read to you and uploaded the PDF it coded all of these and then it gives you a preview. 279 00:30:57,652.0053895 --> 00:30:58,582.0053895 The game itself. 280 00:30:59,242.0053895 --> 00:31:07,12.0053895 And then once you are happy with the game, you publish it, right? But I was looking at this thing and I forgot to actually cite the source. 281 00:31:07,252.0053895 --> 00:31:15,742.0053895 So what I did was I prompted it a little more and said, please add a source citation at the bottom 2025 State of Marketing AI report. 282 00:31:16,552.0053895 --> 00:31:18,652.0053895 And then it recoded. 283 00:31:18,732.0053895 --> 00:31:19,992.0053895 The source is indicated. 284 00:31:19,992.0053895 --> 00:31:21,492.0053895 So now I'm really happy with this. 285 00:31:21,492.0053895 --> 00:31:26,242.0053895 So I published this and then I copy the link and put it in here. 286 00:31:26,242.0053895 --> 00:31:29,242.0053895 And now here is the Jeopardy game. 287 00:31:29,512.0053895 --> 00:31:33,262.0053895 So you can choose any one of these and let's see if we can. 288 00:31:33,712.0053895 --> 00:31:38,452.0053895 You know what percentage of respondents or in the experimentation phase? I have no idea. 289 00:31:38,452.0053895 --> 00:31:40,372.0053895 So I'm going to say 45. 290 00:31:40,492.0053895 --> 00:31:41,332.0053895 Oops, I'm wrong. 291 00:31:41,332.0053895 --> 00:31:44,282.0053895 It's 40, right? Company policies. 292 00:31:44,822.0053895 --> 00:31:49,812.0053895 What percentage of companies lack an AI roadmap strategy? I think that's gonna be pretty high. 293 00:31:49,872.0053895 --> 00:31:50,472.0053895 78. 294 00:31:50,532.0053895 --> 00:31:51,102.0053895 I'm still wrong. 295 00:31:51,102.0053895 --> 00:31:51,372.0053895 75. 296 00:31:54,462.0053895 --> 00:31:59,682.0053895 Like at what company size does Chad GPT usage peak before declining? Oh, I don't know. 297 00:31:59,732.0053895 --> 00:32:00,512.0053895 Interesting. 298 00:32:00,902.0053895 --> 00:32:01,142.0053895 I know. 299 00:32:01,922.0053895 --> 00:32:02,732.0053895 Oh, I'm wrong too. 300 00:32:04,382.0053895 --> 00:32:06,122.0053895 Doing bad this game. 301 00:32:06,272.0053895 --> 00:32:13,142.0053895 But anyway, you can see that my score is minus seven four, the 25 questions. 302 00:32:13,142.0053895 --> 00:32:15,342.0053895 So it's just something fun to do. 303 00:32:15,342.0053895 --> 00:32:17,322.0053895 And the point of this is. 304 00:32:17,742.0053895 --> 00:32:19,182.0053895 I didn't have to code. 305 00:32:19,542.0053895 --> 00:32:27,132.0053895 This is what some are calling vibe coding, which is essentially code word for describe what you want and AI builds it for you. 306 00:32:27,947.0053895 --> 00:32:30,792.0053895 So Ken Jennings would be proud. 307 00:32:31,152.0053895 --> 00:32:32,82.0053895 That was really cool. 308 00:32:32,82.0053895 --> 00:32:37,182.0053895 I appreciate you walking us through what the output is because that's what really matters. 309 00:32:37,232.0053895 --> 00:32:41,292.0053895 For most people who are building it, it's great, but like what the output that's so cool. 310 00:32:41,292.0053895 --> 00:32:45,642.0053895 And that's something that someone could use it in marketing, they could use it in sales enablement, they could use it in training. 311 00:32:45,642.0053895 --> 00:32:46,722.0053895 There's so many options. 312 00:32:46,812.0053895 --> 00:32:47,112.0053895 Yeah. 313 00:32:47,117.0053895 --> 00:32:56,172.0053895 I use it quite a bit for ROI calculators, right? Instead of using a spreadsheet, build an ROI calculator, I use it for dashboards. 314 00:32:56,877.0053895 --> 00:32:58,947.0053895 Marketing op or CMOs. 315 00:32:58,947.0053895 --> 00:33:02,127.0053895 We have so many spreadsheets, turn it into a dashboard. 316 00:33:02,467.0053895 --> 00:33:11,517.0053895 Event planning, put your event plan into an interactive infographic or webpage, right? So there, there are just the possibilities are endless. 317 00:33:11,517.0053895 --> 00:33:17,877.0053895 And I think what the beauty of this is, we no longer have to sit in a, sit on ideas. 318 00:33:17,907.0053895 --> 00:33:23,907.0053895 We can make it come to life, right? Like I have a lot of ideas in my head, but it's too hard to explain. 319 00:33:23,907.0053895 --> 00:33:28,317.0053895 But if I can show it in an interactive way, in a more tangible way. 320 00:33:28,352.0053895 --> 00:33:34,172.0053895 Create a jeopardy game out of it, then it's more engaging and you get to align people faster. 321 00:33:34,172.0053895 --> 00:33:38,12.0053895 You get your points across, hopefully generate some revenue more quickly too. 322 00:33:38,912.0053895 --> 00:33:45,782.0053895 I wanna switch over to the human aspect of work, beyond the ability to prompt what human skills. 323 00:33:46,737.0053895 --> 00:33:53,397.0053895 Emotional intelligence, curiosity, critical thinking will become even more valuable as we mature in our AI usage. 324 00:33:53,787.0053895 --> 00:33:54,117.0053895 Yeah. 325 00:33:54,117.0053895 --> 00:34:06,682.0053895 So all those, like what you just mentioned, right? Like I've always said, as AI democratizes, iq, EQ becomes increasingly important, right? To your point, like the critical thinking, empathy and all those things, are important. 326 00:34:06,742.0053895 --> 00:34:10,972.0053895 But I believe as marketers, balance is actually more important. 327 00:34:11,617.0053895 --> 00:34:18,337.0053895 Because what we are going to need to do is we need to balance innovation with ethics. 328 00:34:19,252.0053895 --> 00:34:26,822.0053895 Automation with the human touch, personalization with transparency, we have to look at both sides. 329 00:34:26,852.0053895 --> 00:34:32,82.0053895 We can't be all human because we forget about all the benefits that we get out of using ai. 330 00:34:32,82.0053895 --> 00:34:34,992.0053895 That we can't be all AI because we forget about the human. 331 00:34:35,232.0053895 --> 00:34:41,507.0053895 So this whole balance thing I actually don't know how well AI would do with balance. 332 00:34:41,867.0053895 --> 00:34:51,917.0053895 We talk about the human aspects, but I'm like, okay, ai, can you actually balance, which is another dimension that it's a lot harder for an AI to do. 333 00:34:51,917.0053895 --> 00:34:53,417.0053895 In my opinion, at least today. 334 00:34:53,657.0053895 --> 00:34:55,277.0053895 I never know what's gonna happen in the future. 335 00:34:55,277.0053895 --> 00:34:56,687.0053895 It's advancing so quickly. 336 00:34:56,687.0053895 --> 00:35:00,717.0053895 So I was gonna say let's keep squinting into the future a little bit and. 337 00:35:01,92.0053895 --> 00:35:14,582.0053895 Of all the go-to-market functions that exist, what new role is gonna be headlining the LinkedIn job boards and which roles are gonna go extinct? There is such a good article in the New York Times. 338 00:35:14,582.0053895 --> 00:35:19,12.0053895 It came out last week, and I just listened to Notebook LM on that one. 339 00:35:19,222.0053895 --> 00:35:25,542.0053895 So I would recommend reading that article, but here's my 2 cents, jobs that guide ai. 340 00:35:27,562.0053895 --> 00:35:52,112.0053895 Because now we're going to have AI teammates and who is going to orchestrate, who is going to direct them? Who is going to guide them with a moral compass? Who is going to say that the work is good or not good? Who is going to put their stamp of approval so that it holds up in court, right? Like those jobs are going to rise. 341 00:35:52,472.0053895 --> 00:35:59,792.0053895 The ones that are, I've always said 60% of the jobs that we have today did not exist in 1940. 342 00:36:00,42.0053895 --> 00:36:04,992.0053895 There were no software developers, social media managers or web designers back then. 343 00:36:05,242.0053895 --> 00:36:09,682.0053895 At the same time, we no longer have elevator operators and we no longer have St. 344 00:36:09,952.0053895 --> 00:36:11,2.0053895 Stenographers. 345 00:36:11,102.0053895 --> 00:36:25,702.0053895 There, there will be job replacements and the ones that are repeatable, the I, one of my good friends who's a language editor her job is gonna go away, she basically said, I'm out. 346 00:36:26,112.0053895 --> 00:36:31,142.0053895 And guess what she's doing now? She's coaching people with English as second language. 347 00:36:31,742.0053895 --> 00:36:31,832.0053895 Boom. 348 00:36:33,347.0053895 --> 00:36:38,697.0053895 That is gonna be a lot harder for AI to replace because it's about, understanding the culture. 349 00:36:38,697.0053895 --> 00:36:44,937.0053895 It's not just the lang nuances of language and culture rather than just editing. 350 00:36:46,17.0053895 --> 00:36:58,737.0053895 English words, being able to transition from something that is fairly mechanical to something that has emotional and cultural political value is something that people will need to think about. 351 00:36:58,797.0053895 --> 00:37:03,742.0053895 Alright, Liza, you probably remember from last year, we close out with some rapid fire quick questions. 352 00:37:03,867.0053895 --> 00:37:06,207.0053895 So you ready? Fill in the blank. 353 00:37:06,267.0053895 --> 00:37:11,187.0053895 AI is powerful, but it can't feel. 354 00:37:12,87.0053895 --> 00:37:18,477.0053895 What is one AI tool you can't live without? Liza, GPT. 355 00:37:19,197.0053895 --> 00:37:19,677.0053895 Perfect. 356 00:37:19,767.0053895 --> 00:37:21,267.0053895 One prong, a digital twin. 357 00:37:21,627.0053895 --> 00:37:22,77.0053895 I love it. 358 00:37:22,77.0053895 --> 00:37:22,647.0053895 That's perfect. 359 00:37:23,247.0053895 --> 00:37:25,587.0053895 What a role You we just talked about that. 360 00:37:26,397.0053895 --> 00:37:33,57.0053895 If you could instantly upgrade every go-to market professional on one skill, what would it be? Balance. 361 00:37:34,347.0053895 --> 00:37:34,737.0053895 Love it. 362 00:37:35,377.0053895 --> 00:37:36,277.0053895 That was awesome. 363 00:37:36,447.0053895 --> 00:37:38,517.0053895 Thank you so much, Liza, for joining us. 364 00:37:38,517.0053895 --> 00:37:40,167.0053895 It's always great to have you. 365 00:37:40,267.0053895 --> 00:37:42,517.0053895 We really appreciate it and I know our listeners do too. 366 00:37:42,522.0053895 --> 00:37:42,912.0053895 Thank you. 367 00:37:43,362.0053895 --> 00:37:45,392.0053895 Thank you so much for having me. 368 00:37:45,572.0053895 --> 00:37:46,472.0053895 You guys are awesome. 369 00:37:46,962.0053895 --> 00:37:47,432.0053895 Thank you. 370 00:37:47,437.0053895 --> 00:37:48,537.0053895 Thanks, Liza. 371 00:37:49,237.0053895 --> 00:37:49,897.0053895 We'll be right back. 372 00:37:50,737.0053895 --> 00:37:55,232.6911038 uh, uh, uh Bye. 373 00:37:55,232.6911038 --> 00:37:56,932.6911038 Another fun conversation with Liza. 374 00:37:57,242.6911038 --> 00:38:05,132.6911038 Ken, what was your biggest takeaway? The biggest takeaway from that conversation was at the very end when she was talking about what skill. 375 00:38:05,882.6911038 --> 00:38:09,812.6911038 She wished every go-to-market professional had right now, and she said balance. 376 00:38:10,112.6911038 --> 00:38:19,412.6911038 And with AI moving so quickly, you gotta keep focused on the human aspect and a little bit of self-regulation to not overwhelm yourself. 377 00:38:19,562.6911038 --> 00:38:23,547.6911038 So I do think being balance is actually going to be a skill that people. 378 00:38:25,817.6911038 --> 00:38:28,907.6911038 Explicitly plan to develop over the next couple years. 379 00:38:29,197.6911038 --> 00:38:32,647.6911038 And I also think balance from a perspective of how you lead a business. 380 00:38:32,647.6911038 --> 00:38:41,157.6911038 We've talked about this before, but the idea of Jack and Jills, of all trades this is maybe their era to thrive because they have a skillset that's very balanced. 381 00:38:41,277.6911038 --> 00:38:45,937.6911038 So I think balance is going to be the theme of, the next five years. 382 00:38:45,937.6911038 --> 00:38:46,837.6911038 That's my mindset. 383 00:38:47,437.6911038 --> 00:38:48,127.6911038 Oh, I like that. 384 00:38:48,367.6911038 --> 00:38:48,577.6911038 Yeah. 385 00:38:48,787.6911038 --> 00:38:55,777.6911038 What about you? What's your takeaway? I think the Gladiator is one of my favorite segments that we've launched this season. 386 00:38:55,857.6911038 --> 00:39:09,277.6911038 And I think making it fun and the Jeopardy game was a prime example of something that, could be dense to read through with having a really dense research report that you want people to be aware of. 387 00:39:09,517.6911038 --> 00:39:14,957.6911038 But being able to make it fun and being able to do that without an engineer and create a little game is. 388 00:39:15,247.6911038 --> 00:39:15,787.6911038 Pretty neat. 389 00:39:15,787.6911038 --> 00:39:22,857.6911038 I'm a big fan of the creativity as I mentioned, Liza was the one that turned me onto the Claude a Artifacts, and I've had a lot of fun playing with them. 390 00:39:22,857.6911038 --> 00:39:25,367.6911038 But I think that, there's so much more potential there. 391 00:39:26,777.6911038 --> 00:39:27,647.6911038 Yeah, I agree. 392 00:39:27,647.6911038 --> 00:39:33,827.6911038 I think making AI fun is also how people will stay balanced and sane when doing all of this. 393 00:39:34,7.6911038 --> 00:39:35,767.6911038 And I'm gonna try to make like a game. 394 00:39:35,767.6911038 --> 00:39:37,627.6911038 I don't know for what use yet, but. 395 00:39:38,252.6911038 --> 00:39:39,92.6911038 I'll figure something out. 396 00:39:39,92.6911038 --> 00:39:39,422.6911038 I'm sure. 397 00:39:39,602.6911038 --> 00:39:40,52.6911038 Hey, I like it. 398 00:39:40,52.6911038 --> 00:39:42,542.6911038 Let's put balance on our Bingo card for this year. 399 00:39:43,502.6911038 --> 00:39:44,72.6911038 I love that. 400 00:39:44,562.6911038 --> 00:39:46,392.6911038 It's been another great episode. 401 00:39:46,392.6911038 --> 00:39:52,992.6911038 Special thanks to our friend of the pod Liza Adams, for sharing all of her great insight. 402 00:39:53,92.6911038 --> 00:39:55,402.6911038 And thank you for tuning into Futurecraft. 403 00:39:55,402.6911038 --> 00:39:58,912.6911038 Go to Market if you wanna stay ahead of what's going on in AI and go to market. 404 00:39:59,602.6911038 --> 00:40:05,2.6911038 Give us a follow, if you really like us, we'd love if you could give us a review and leave a comment, share with a friend. 405 00:40:05,212.6911038 --> 00:40:07,102.6911038 This is how we get our voice out there. 406 00:40:07,522.6911038 --> 00:40:09,292.6911038 And thanks for listening. 407 00:40:10,12.6911038 --> 00:40:10,402.6911038 Yeah. 408 00:40:10,672.6911038 --> 00:40:15,472.6911038 And until next time, let's keep crafting the future of AI and go to market together. 409 00:40:15,772.6911038 --> 00:40:16,312.6911038 Thanks.