The AI-Augmented CMO: Architecting Teamwide Adoption

[00:00:00] Hello, Renegade marketers. If this is your first time listening, welcome. If you're a regular listener, welcome back.

[00:00:07] You're about to listen to an expert Huddle where experts share their insights into topics of critical importance to our flocking awesome CMO Huddles community. Recorded back on April 16th, this episode features Francois Dafoe on what it takes to be an AI-augmented CMO. I think many of the things he talks about are still relevant.

[00:00:30] One thing is interesting, after this recording, Francois joined Anthropic, which makes revisiting this one even more fun. Francois argues that AI transformation isn't something leaders can outsource or admire from a distance. It starts with getting hands-on, making time to build with your team, and giving AI enough context to do more than save time.

[00:00:55] For Francois, that is when things start to get really interesting, and I think it will be the same for you. If you like what you hear, please subscribe to the podcast and leave a review. You'll be supporting our quest to be the number one B2B marketing podcast. All right, let's dive in

[00:01:13] Welcome to Renegade Marketers Unite, possibly the best weekly podcast for CMOs and everyone else looking for innovative ways to transform their brand drive, demand, and just plain cut through proving that B2B does not mean boring to business. Here's your host and chief marketing Renegade Drew Nier.

[00:01:43] Hello hustlers. Today we're diving into what it really means to be an a AI augmented CMO with someone who's living it every day. Francois dufour has LED marketing at companies like Twilio and Algolia, and now operates as A CMO Wingman supporting multiple B2B leaders at once, using AI as a force multiplier.

[00:02:05] This conversation builds on what he's. Seeing across dev tools, cybersecurity and AI native companies, and what actually changes when AI becomes part of your team. And I know we're all talking about this, but Francois is doing it. So Hello Francois, how are you? And where are you This fine day.

[00:02:22] Hey Drew. Uh, I'm in home in Palo Alto today.

[00:02:26] Thanks for having me. I look forward to discussing this.

[00:02:29] Yep. I think we're gonna have to have you at our CMO super huddle in October because this conversation is not gonna end here. But let's do what we do, uh, in most of our expert huddles, which is just in case our audience needs to leave early. Or we need to convince them to stay.

[00:02:43] Please list three mistakes you see marketing leaders make when it comes to becoming ai, augmented CMOs. And now just list them. Try it's really hard. I know. And then we'll go through 'em one at a time.

[00:02:57] Sounds good. the first one's gonna be obvious. It's delegating the AI work to their team and not being hands on themselves.

[00:03:03] Okay. That's interesting. And we'll go. Uh, so not being hands on. Okay. Number two.

[00:03:08] The second one is not blocking the time to build stuff collectively and repeatedly. Okay. Because doing it together is, it changes everything and doing it again and again. One and done doesn't work, is not enough.

[00:03:20] Okay.

[00:03:20] And the third one, it's being too nice to their team, not daring to publicly shame the people who don't use, uh, and push AI enough.

[00:03:28] It's really great to do leaderboards and be positive. but at some point you sadly will, may need to let people go if they don't wanna lose ai. So shaming them publicly is actually doing them a service. And most CMOs I see, don't dare do that.

[00:03:41] Yes. Executives

[00:03:42] in general.

[00:03:43] Alright, well let's go through these one at a time because delegating is really interesting.

[00:03:48] You know, we just got back from three, strategy labs in San Francisco, Palo Alto, and Seattle, and a lot of the CMOs and I, and I thought it was great actually. Were really. Spending weekends getting up to speed on what, what the art of the possible was. And so they were putting a lot of time into, you know, whether it's vibe, coding or building stuff themselves.

[00:04:07] So how is it that yes, they need to delegate, but they also need to kind of know what the art of the possible is, right?

[00:04:14] Yeah, absolutely. So that, that's why it's important to stay in touch with what's going on, what the others are doing. Exactly. Joining the sessions like this get inspired. So understand both what's possible, what.

[00:04:26] What others are doing, and I'm saying you can't delegate entirely to your team because you need to be acting as the top architect, just like you are the one deciding what context, what direction I'm giving my team. It's the same thing with ai. You need to actually, just like you onboard new employees, you need to onboard the agents.

[00:04:42] You need to onboard ai, give it the right context and architecture, all this. The CMOs I see doing that very well, are deeply involved in making sure that's happening.

[00:04:52] So they are making sure, and you, you call it architecture, I could just call it foundational work. Uh, no different than doing foundational, strategy or positioning work.

[00:05:02] You need to do the same thing with AI at what stage? So we've set the strategy or the architecture and we've provided all the context. what are we delegating? I mean. All the bills, all the use case. What, what's, what's happening there that you're seeing that's working?

[00:05:18] Oh, so some people, some CMOs are delegating the building, the, the central context and central intelligence layer, typically to someone in product marketing and someone in corporate marketing. Make sure that's all centralized, all those are delegating the, of course, what you'll be doing on the growth side, go to market engineering.

[00:05:35] creating the campaigns and the systems, more critic. Critically, it's almost the architecture of how does, do we do the integrations between or wherever our knowledge lives, whether that's notion or other places like this, and how do we give access to our, the AI or Theis that we pick? Uh, when I'm saying they cannot delegate everything, for instance, when they do hackathons, typically CMOs, please don't be a judge, be a player, because they, it is tempting to be, yeah, I'll be a jury, I'll be a judge, and I'll see what the teams come up with.

[00:06:02] And, and give them feedback. It sets you. You need to be hands on, understand what's going on, especially what's missing. but also it shows the, the team you're leading by example, and most critically, the strategic leverage that using AI as a partner gives you as a CMO is really huge because we know how to guide it strategically and we can get.

[00:06:22] Leverage from it that shapes our strategies, our directions that typically an IC on the team wouldn't, be able to see with the same level.

[00:06:29] That's a amazing, and I, I, I guarantee you, I know a number of CMOs in our community have done hackathons, but I guarantee you they have not been participants. So I love that notion, in number one and number two in theory, if what you're saying is because you understand the context, because you understand that you will be able to be a huge participant 'cause you know how to leverage, you know, you're, you're coming in.

[00:06:54] In a sense, one step ahead. 'cause you have a firmer grasp of the strategy is did I add two things together that weren't there? Or was that what you were saying?

[00:07:03] No, that that sounds right. at, at both levels. So 30,000 feet, you see strategically what the agents and the teams have access to and what's missing context, OKRs, construct goals, et et cetera.

[00:07:13] and also technically you see. Where they're hitting some issues. Like this morning, just I was with a team and you realize they're trying to connect one of their tools to child DPT and oh, they don't have access to this. The CSO has not accepted that. The fact that the CMO catches that quickly, and by the way, the CEO was in the room at that point, boom, they can actually go prioritize and make that happen.

[00:07:35] Interesting. Okay. So the second one that, mistake that you're seeing, and I'm certain that this is true, is they're not building the stuff together. And talk about what you mean by that and what that looks like when it's working well.

[00:07:47] Yeah. Well first, the, the first thing is the CMOs I work with rightly so, are so busy.

[00:07:53] Everybody's so busy. And so, finding the time to do this In their reit, this is, they can actually decide we're gonna be doing these things at upsides, hackathons, et cetera. And it's for them really to help the teams block the time so that together you cannot keep having the excuse, I'm too busy to learn the things that will make me less busy.

[00:08:12] and so it's blocking times, for instance, the all hands your cm, all hands have every time two to three people who show their cool and latest workflows. It's actually not thinking I'm gonna do a AI hackathon once a quarter and then we're done. Uh, there are better models for that. Um, an AI trainer yesterday I was talking to has been doing that frequently thought.

[00:08:31] Yeah, the initial training is good. What actually works very, very well. For instance, when you do sprints with a small team of, within three days, you do three, uh, 90 minute sessions with a small team and they build something. Uh, so you create urgency and you, you keep doing this together because. Otherwise there's context and priority set in and they take over.

[00:08:52] Well, and I, I think the, the notion of the CMO showing up in one of these, uh, scenarios that you're describing is sort of is walking the walk and saying by being there, the medium is the message. Because you were there. This is a top priority for you that the organization is gonna transform using these tools.

[00:09:10] And that's why I'm here and making time for it. It's hard. It's hard, but what, what I, I think we're behind this that we haven't really stated is. This ultimately is going to radically change the productivity of your marketing organization and hopefully you. And so being in there, uh, in some of these builds is, is really important.

[00:09:32] Okay. And now the third one. Being too nice. Now I wanna push back just a quick second because I saw the extreme of this, and there's a lot of research on this, and a friend, uh, CMO was at a panel recently where folks were saying, yeah, it's, it's gonna cause some job loss. And, you know, bummer about that.

[00:09:50] And the reality is most. companies, the CMO surveys show that have a very small budget for training and that budget just went down this year. So they are not training their people the way they think they are training their people. So maybe, and you have a story about that. I think you mentioned meta yesterday, but, I don't know if everybody's training and so therefore, but they have the expectations that everybody on their own, on the weekends, whenever is learning these tools.

[00:10:17] So before we get to. Hey, see ya, you're not using the tools. Isn't there some important training that has to be in place?

[00:10:27] Uh, yes. So it's important enablement and so the, the training you can bring in trainers and now of course things change and evolve so fast that the, the trainings keep evolving and that's why it's important to learn together because you are figure out things together and the, the story you're referring to, meta.

[00:10:45] Recently Zuck asked a lot of his product and Edge teams to do nothing for a whole week, but learn how to use ai, set up their assistant, set up mostly their agents. Even when, so far as, uh, doing, you can even postpone for a week, the, that corresponding week launches, and projects because it's critical.

[00:11:03] We do this together at the same time. Now, they, they prepare, they prepared for this. What they did is they had their, their teams, IT and probably the platform engineering teams. Set up the equivalent of open Cloud called MyCloud that they set up on their own compute. So it was secure, it was accessible.

[00:11:17] They set up what they call Second Brain, which is a place where everybody can also have their, their knowledge and context. And for a whole week people were said, don't worry about cost, you just go do this. what happened is they were tracking. Which people were doing the most, using the most tokens, et cetera.

[00:11:32] There's the trend called token maxing, where people are just trying to show up, especially on these leaders board to leader boards to show they're using a lot, a lot, but it's a good sign. So they were tracking also who's not using it, who's not billing anything good with ai. And sadly, you've probably seen the rumors.

[00:11:46] They're about to cut a lot of people, and I'm pretty sure that's been taken into account in terms of who stays in who, who, who leaves. I've talked to a friend of mine for a week actually. He set up his own agents. And working with this, they were pretty much not on their own on how to do this, but they figure it out and they learn.

[00:12:02] And at the end of the week, he was so happy of how much of the stuff he hates, that he automated these, um, satisfaction with jobs seemed to be much greater and also understand what's possible. but at some point, if that's not enough, then this is where I'm going to shaming. You need to call out people who are not trying.

[00:12:20] And figure out actually really when you call 'em out, figure out what they need. Where are they getting blocked? Is it a mindset problem? Is it an email problem? Is it they, they, their tool is not a proof problem. Just be very empathetic to them. But at some point, also, after doing all the positive stuff, you need to just tell them it's not okay.

[00:12:39] can't argue with you at all. There's no world that I can ima, and you're doing them a favor in the sense that, their next job, they're gonna need this. It's no different than, you know, if you went to a company and said, by the way, I don't, I don't use computers, um, I don't use mobile phones.

[00:12:53] I don't use, that's the way ai, and there's, so there's no future job. Where you won't be using these tools necessarily unless you're ready to be a plumber or a electrician. And, and by the way, those are gonna end up, you know, you're gonna end up using 'em for those too. So, uh, you're doing everybody who's at your company a favor, if you can bring them up to speed.

[00:13:14] I have one issue that I wanna sort of get at, which is w. What you measure matters, and that's what drives people. It's like if you incentivize sales strictly on sales, they will go and get the lowest hanging fruit. Whether or not those are, in your, uh, ICP, whether or not, you know, the metrics that you, that things that you measure matter leaderboards are simply measuring.

[00:13:40] Activity. They're not measuring results, impact efficiency. It's just you used a lot of fricking tokens and so I'm wondering if that's really misguided and there's another way to do this because look, I could spend a hundred percent. Percent of my time on ai, and that would be impressive, except for the fact that I spend six hours a day talking to CMOs, which is a lot more impressive.

[00:14:08] Yes. Right. So that's the problem with token maxing it. It can drive the wrong behavior.

[00:14:13] Right.

[00:14:13] But at least it tracks something and. So it is far from perfect and it should be changed quickly, but at least when, and, and it incentivizes you to run agents on the side and, and yeah. But at at least that means people know how to actually build agents and to the, the effect that you're still managing them by OKRs or whatever goal you and system you're using, they still need to do and perform in their job.

[00:14:35] so quickly you should change. And the, what I like is instead of leaderboards. Well on top of leaderboards, which you can also have and be smart indeed about what, what you feature there. It's having them build or at least, complete and publish a list of their best workflows, their base agents. It could be a very simple Excel, Google spreadsheet or our table, whatever it is.

[00:14:55] And you show them, you ask them to complete what's the workflow? is it something that makes me gain time or provide more strategic value to customers or internally? Roughly have built it and see maybe an estimate of how much time it helps them save and everybody has to publish that. And then you actually take maybe every time you've got an all hands, two to three examples when you have these folks actually showcase their workflows and you share that list and everybody else will go, hopefully at some point, oh, this is cool.

[00:15:20] Can you show me? And so this is really about finding a system. That surfaces your, your top people. 'cause this what this is where, how you will find your champions who's more advanced. Hopefully those who also are on top of being champions are angels willing and to spend the time to educate others. And also a simple tool to educate others.

[00:15:39] Uh. Start with that and, and it makes sense to me. But I wanna go back to something you said earlier, which is this a group sport, a team sport? Yeah. A lot of these workflows, and I'm sort of imagining, wait, uh, the guy over here is solving a workflow that 10 other people at the company have. Couldn't the two of two people from that company work better faster, solve a broader workflow?

[00:16:00] I mean, if everybody's out there solving these problems, are we really of being efficient? so I don't know. I, I mean it there, it feels like as a, as a leader, you, one of your primary task is allocating resources, right? Yeah. You set the vision, you hire the team, and then you allocate resources. So in that process, do you want everyone on your team trying to figure out.

[00:16:24] Better workflows. I don't know. Maybe you do. I'm just, I'm, I I

[00:16:28] think it's a mix of both. It depends on the workflow, whether, first of all, is this workflow applicable to more than one person or more than a few people on the team? How complex is it to build? Does it require some IT help and some dev actually, some people will know both.

[00:16:41] Some vibe coding, but also integrations and, and you, you look at those. Ideally, everybody should be able to set up their own simple workflows using whatever tool they can. But for things that become more systematic, either are the most critical workflows for a given, function. So for instance, for product marketing or for, then you need to find who are my system thinkers,

[00:17:03] right?

[00:17:03] Are they technical enough or at least do they, do they think enough in in systems to actually be the ones that are assigned to go and solve this, collect the requirements. Take the lead on that. It could be someone in PMM, it could be someone in corporate marketing, it could be someone in in GTM marketing, or maybe you bring in some, some other external help, whether that's external the company or from it.

[00:17:23] And this is your role, either as the CMO or you. maybe that's a very good projects for head of marketing ops, slash chief of staff to do that. Also, to your point earlier, drew the. There's a, a best practice is to further enable your champions. If you find people who are really, really good, make sure you give them extra budget, extra time, maybe even their own upsides that you pay for where they get together and solve bigger problems.

[00:17:48] 'cause then they'll come back to the company and further enable.

[00:17:52] Okay. So just to wrap up this segment, we've got three things that everyone here is gonna walk away with. Don't de delegate at all. Make sure you're in the mix there. two, building stuff together, uh, that it's at this team sport and, uh, so do that.

[00:18:06] And yeah, we probably don't wanna be so nice or forgiving about slow adopters, uh, in particular. And then sort of the. The extreme side of that is when you find these system thinkers really support them because you need them and sort of, you know, I think it's easy to predict that every marketing team will have an AI engineer on there next year, um, uh, if not sooner.

[00:18:30] So that's probably the extreme side of that. Okay.

[00:18:33] Two things. It's one, the most critical thing is probably blocking the times together. It doesn't mean you need to do and build all the workflows together, but if you block the same time together, it gives you the license to actually spend the time and there's no, the other slacks happening and everything that's dying for your attention.

[00:18:49] Right.

[00:18:49] And the other things they may need in indeed, things where you need to, match the person who's really good at the marketing craft and the person who's really good at the systems. And ideally you can have some small groups like that.

[00:19:00] Yeah, it's so funny. It's just, it's not that different from the early days of web dev in that you needed some people who could actually, you know, code and knew how to, you know, HTML and then you needed people who knew how to communicate and, and they were often different people, so, uh, Isabel, you have a question? Why don't you come on camera and we'll, we'll just get to it right now.

[00:19:20] Yeah, sorry, uh, for the interruption. Hi Francois. Um.

[00:19:27] From one French person to the other. Uh, could you just repeat, drew, the second point you made, you summarized the three points, the second point you made about the late adopters, because I think what I heard was like, call them out, but also be empathetic to understanding what's blogging them. Is that, did I hear it correctly?

[00:19:43] Yeah. Well, in fact, thank you for jumping in because yes, it was your conversation this in a conversation that stimulated this. So there, I think the, the point is train them so, and if they're still not. Using the tools and adopting it. You have to find out what's blocking that. and sometimes there's a green reasons, sometimes there's legal reasons, whatever it is, whatever that resistance is, um, you need to understand that.

[00:20:10] And then I know one CMO sort of said, let's work on this together. here's a challenge. We can do this together as a, you know, we're gonna build an agent for you together. And walk 'em through it, and if they still don't embrace it, what Francois said is there's no more nice guy. I'm not sure about public shaming.

[00:20:27] I've never seen that work effectively as a leadership tool, but I do think the opposite of celebrating those people. Who are doing it once a week. Hey, here's a cool thing that somebody did in service of the strategy. Again, it can't just be a cool thing someone did. It's gotta be connected to some big program.

[00:20:49] You as the CMO believe, again, I, I feel like. Efficiency is not the goal. Doing something you could never do before is the goal. by the way, there was a great editorial in, uh, uh, economist by Ethan Molik, who is a professor, uh, at Wharton who companies should treat AI as what it is profoundly weird. And he just talked about that so much of energy is going into being efficient when in fact there are, it's a weird tool.

[00:21:17] Do weird things with it, like do new things. So I love that. So did we answer your question, Isabelle?

[00:21:23] Yes. Thank you for the

[00:21:24] clarification.

[00:21:24] Okay, well, we'll talk in a bit.

[00:21:26] We're gonna take a quick break. We'll be right back.

[00:21:30] This show was brought to you by CMO Huddles, the only marketing community dedicated to B2B greatness, and that donates 1% of revenue to the Global Penguin Society. Why? Well, it turns out that B2B CMOs and penguins have a lot in common. Both are highly curious and remarkable problem solvers.

[00:21:51] Both prevail in harsh environments by working together with peers, and both are remarkably mediagenic. And just as a group of penguins is called a huddle, our community of over 300 B2B marketing leaders huddled together to gain confidence, colleagues and coverage. If you're a B2B CMO, why not dive into CM O huddles by registering for our free starter program on CMO huddles.com.

[00:22:21] Hope to see you in a huddle soon.

[00:22:23] So, alright. you use this con word context that the, the hardest part of using AI has nothing to do with the tool. Can you explain, make that practical for CMOs trying to get real leverage?

[00:22:36] Uh, yes. So there's this notion that context is everything in the way you expose your AI to your business customer context. And the the reason is when you think of what's needed is really good models. Harness is actually the system that manages your agents and context. If I simplifies these things first, two, you're not building the AI labs and the tools.

[00:22:58] They're building that and they're shipping extremely fast. Uh, so it's improving so fast that you'll, you'll reap these benefits anyway. It's the third, the context that, is essential because when you give, even with a prompt that's not great, the right context to these really good models, you get good outcomes.

[00:23:16] And the third part of that context, the AI cannot invent it. They cannot come up with it. It's, this is exactly what, uh, as the CMO is, you're, you're leading the team setting vision, direction, priorities, et cetera. And so these agents systems are becoming so good that what we need to spend more time on is giving them access to this well-structured context.

[00:23:35] That can be your brand style guide. Persona guides, product messaging skills with and processes that are unique to your team. When you've got that, then the system can actually pull up both orchestrator and Subagent without us even knowing and working on those. That's why that really matters.

[00:23:51] So give an example.

[00:23:53] Real world example of, I'm just thinking, uh you know, the simplest one is you have your brand agent. Uh, the AI generates some copy. That copy goes through the, then the brand agent to make sure that it's on brand. Then you might have another agent that says, Hey, stylized, why is this needs to be? A lot better.

[00:24:12] I don't know. What, what's a process that, you know, you have context when that, and, 'cause I know you have real world examples.

[00:24:21] Yeah. So, uh, I'll talk if you want my, my personal case, what I do is I advise eight to nine different CMOs at a given time every week. So I've got a lot of context switching. And that's probably the same for all of you.

[00:24:31] You got different projects. Different products, different uh, stakeholders, and so I spend a lot of time consolidating my contacts. I use granola for every meeting. I've organized everything around folders. I've got one folder per client. Some other ones on top. I've created the same exact folder structure in claude, which is my go-to, uh, ai, the same exact folder structure in Google, drive because my workspace is also where I got my knowledge and I make sure that at the end of the day I created a little script that extracts from granola all the, the summaries of my chats, my meetings, and puts that into.

[00:25:10] Google Drive, Google my Google Docs. So first of all, my clients then also have access to these notes, but also it keeps collecting that context, uh, so much. so that's number one. Number two, then what you're describing, these agents, for instance, if you want to have a bread agent, then that means you, you, whether you create a scale, you need to actually give it all, all that information and that information needs to be updated.

[00:25:33] I've got not only tools that at least all of my, my folders, which collect all that history, but also some skills that I've sometimes borrowed from AI trainer and friends that also, um. Uh, customize and then you give access also to that, to, to claude.

[00:25:49] So can you give an a specific example of a skill that you're of applying to your setup?

[00:25:55] Yeah. So for instance, the first, it, it's, um, a skill that is really good at using existing positioning and narrative frameworks and applying them to the context that I've said about my clients or other input. So, for instance, and you're all probably familiar with anti raskin strategic narrative. With April Dun Ford's positioning, I've created a skill that actually did some, first of all, some deep research online coming up with, the best frameworks to do that.

[00:26:19] And then I, I give it an input, which is, Hey, look at this landing page. Look at this sales pitch. Look at this post, and using the, I call it the Narrative analyzer Scale determine, first of all, which of these templates or frameworks is the right one to use. take two of them and apply these frameworks to critique the good and the bad and how this story can be improved.

[00:26:39] So that, again, using, first of all, input scale and, and then something that can repeat and tweak.

[00:26:46] And so the output of that is, material that you can then talk to your clients with? Uh,

[00:26:53] yeah. Or that I, just looking for myself, it's basically gonna. Tell me, well, I use this framework and then indeed, this is really what's clear, what's working, this is what they can improve.

[00:27:02] and it just gives you something. Of course, I never send that as it is to, to my clients. Right. But it just gives me, at least informs my opinion to actually then have a discussion with them and tell them, yeah, I think we, we can improve there. Or we, we could actually double down on this.

[00:27:15] Okay. Um, we do have another question that we might as well.

[00:27:18] I mentioned this notion of an AI engineer. Maybe you could talk about if someone, what is this person? I mean, we've talked about system thinking, but how, uh, do. There the question, what are the proof points that they actually know how to do this and skills that they have. Uh, and I don't have any idea if you know what the salaries are.

[00:27:41] Well, salaries, I'm, I'm gonna get, well, it's gonna vary Great. Greatly by location, seniority, I'm sure. Uh, the easy way to think I think about this. It's basically the modern AI version of marketing ops, right? But if you think of marketing ops, there's the people who run the campaigns, the people who do analysis, and the people who set up and build stuff.

[00:27:58] It's more the latter. It's the builders,

[00:28:00] right,

[00:28:00] who also understand how to use and, and bring together different tools. And they are AI first, the, the way you figure out they're good or not, you simply ask them, Hey, show me what you've set up.

[00:28:10] and how did you get, first of all the requirements. How did you get the idea?

[00:28:13] and, and then how did you evolve? Um. The workflow or the agent, and you ask them to demonstrate, and you're listening for, is it creative? Are, is it, is it teaching me something? Is it, are they technically pretty good? Are they using the right modern tools? And then, you know they're onto something,

[00:28:29] Yeah, I mean, and the difference between this person and me, uh, and like I can go in and write the best prompts in the world and find them and borrow them and put them in there. I, at the moment, don't really know how to connect an API, and I know I could probably go into Notion and figure it out, but I don't want to, it's not my thing.

[00:28:46] I'm not that kind of thinker. And so I want people who wanna do that, who wanna go in and connect those, those things. I do think that some of this is more than AI usage. This is understanding how technology can work together.

[00:29:01] Yeah. I think they, they need to understand some of these building blocks like that.

[00:29:03] Indeed. So for instance, my, uh, I'm not a developer, but I used to run marketing at Twilio. So I do understand APIs. but when I use, uh, claude code for instance, a lot, very often I need to actually go to a specific tool and get the API key, come back in. I understand the difference between an MCP connector and an API, when to use that.

[00:29:21] And this is, you need to find someone who, who gets these differences.

[00:29:25] I don't know what the difference is between an MCP connector and an API. I don't know if the other folks listening are. Could you just give us that definitions for a second?

[00:29:34] Yeah. So the API Application programming interface is basically the language through which software communicates together.

[00:29:40] Right. This is really what's been used and mostly used by developers to actually make two different tools work together. It's robust, it's safe, it's secure. It requires some coding to assemble. And, um, the MCP think of that in an abstraction layer just on top. So the MCP. Now when you go to claude, for instance, they call them connectors.

[00:29:58] It's an abstraction layer on the top of the API. It's a little simpler. It's not as secure. It's usually not as fully featured and not necessarily as performance, but it's much easier to use. Okay, and, and they've got, yeah, limited functionality, but if you go into claude Core for instance today, and you want to try one that's really, really easy to use and fun, you go to connectors, you, you enable a few of them, they have Google Drive.

[00:30:21] Gmail, Google Calendar. My favorite is granola. And you see connectors. You can actually then, uh, add them. You'll need to authenticate yourself. Of course you get many other ones. Gamma, you get more and more and more the, then you go to, you just pick one and go into the setting to see. The, permissions you can give them.

[00:30:40] And then you see all the capabilities. So for instance, in Gmail, it can read your email, it can create drafts, it cannot necessarily send, right? So you've got limited functionalities in as mcp, again, much easier to use, et cetera. But when you see the permissions, you get a really good understanding what they can do.

[00:30:56] And sometimes this is where you realize, ah, shoot, the stuff I really wanna do is not doable with the MCP or the connector today. And then you need to actually get into maybe some vibe coding using claude Code or Codex, or find someone who does that for you.

[00:31:11] And this, your AI engineer should be able to do that for you.

[00:31:14] And again, if we're delegating stuff, I don't necessarily think the CMOs are gonna get in that level, but they need to understand a little bit about how these things come together. So, this is a kind of a weird question, but you talk about, uh, how this is all integrating with Gmail and G Drive and so forth.

[00:31:30] A lot of us live in the Microsoft Outlook world. Is, are we crippled because of that?

[00:31:36] I told you Microsoft Outlook, uh, world, I, I think you don't have the same capabilities necessarily. Now. Uh, for instance, Claude doesn't have the best connectors to Google. Uh, Claude has amazing, or at least it depends on what's Google building, the MCP connectors,

[00:31:50] right?

[00:31:50] Great connectors. Granola has a great connector. granola and notion have great connectors. So it, it, it's evolving very quickly. and so you need to just, um, figure out what your team uses as your IT team or just ask your, um. More technical people to tell you. However, going back to your earlier point, if you're a CMO, uh, you, you should understand how connectors work and which ones you have access to, because they're the tools you already are using, right?

[00:32:14] Including you using Slack. This is Slack connector. You are probably using calendar and email. Well, you are. I know. Uh, so go figure out what's in there. 'cause cowork, in the case of cloud or codec, it needs to also become your, your thinking and strategic partner. An example that I uh, figured out this week, it's so, I can't believe I didn't see it earlier.

[00:32:32] You go in claude cowork and there's a, a set of playbooks they have. And if you use, uh, granola for instance, and Google, uh, drive, you can simply enable these connectors and create your daily brief. And every day is gonna run a task every day is gonna create very easily a little digest that shows you who you're meeting, the critical emails you need to pay attention to.

[00:32:53] and, and a few points about your day in the, just two prompt. Oh. Uh, the prompt is preset. Just enable the connectors and later run.

[00:33:01] Sounds like this is something that everybody should have. I certainly don't have my daily brief. I am not using cowork. I will, the question I sort of have on that just in general is how, again, do I have to switch our company off of Outlook, uh, in, in order to really do this?

[00:33:19] We can probably do it in Outlook right now.

[00:33:21] There, there's gotta be some ways that Outlook works with, uh, well, charge GPT Cloud and others especially. For instance, cloud is really, really good with the Excel three, uh, the, the MO with Microsoft 365 Suite. So it's, it's got great integrations into PowerPoint, into Excel, and it just release also work.

[00:33:37] Okay. So fine. Well, uh, to be determined, uh, I do feel like pilot, you know, co-pilot on Microsoft has limited a lot of folks that are just Microsoft soft shops where they can't use Claude. I feel very bad for them. but I also feel like every CMO who's in a Microsoft shop better befriend their IT and their, their lawyers and get over that because they are gonna fall behind.

[00:34:01] I mean,

[00:34:01] and

[00:34:02] then I just saw this morning, there's a connector between Chat GPT and SharePoint. So if you're a Microsoft SharePoint, there's, there's a connector and then there's also tools like Lean that also help you consolidate all that context.

[00:34:13] if your company will let you use cloud, if your company, and there are companies out there yeah.

[00:34:18] That are not letting their employees use TBT or Quad, they can only use what Microsoft has. I feel bad for them. By the way, side note, if you know a guy, somebody who's killing it with copilot, please let me know. 'cause we gotta get them here to help those folks that are stuck in Microsoft. Uh, and because it's a problem for them, but they gotta, we gotta help them, probably more than the others.

[00:34:39] Alright. You, let's see. We've covered a lot. We've talked about granola and Claude and how you've sort of got your daily brief. Um, you're also having AI surface recommended topics for each client. Is that part of your daily brief?

[00:34:53] Yes, it is. Uh, okay. So it is simply thanks to this great granola integration where I've got all my meetings organized by client folders.

[00:35:02] I simply, it was one prong. I, I told, I told Claude. I think it was cowork, to tell me for every meeting that I have, when you see there's something coming up on my calendar with a given client since I, it, it came up by analyzing, uh, my gran meetings with who's my client. You could guess. So now he's got this.

[00:35:21] this list of my clients map by companies. it's identifying who's, uh, I'm talking to. It's, I've told it. Look through the last three meetings or at least the last meeting, tell me what are the six things or up to six things they talk about. We talked about and recommend three to five topics we should tackle today.

[00:35:36] for those who have not used granola, including myself, so are you taking your meetings on Zoom and just inviting granola? What's happening? How is granola getting your into your meetings or are you, what functionality does granola bring to the table?

[00:35:49] it runs locally on my Mac. It's running right now, but you don't see it in the Zoom.

[00:35:53] So it's transcribing everything we're doing. It's not recording. . It can, it automatically generates summaries. I can structure the notes the way I want as I'm, um, I'm working. And you can create the equivalent of, uh, it's called recipes, granola food theme.

[00:36:07] Yeah. Yeah. You can create at the end, just like within an AI notetaker, you can just ask any question from it. But also you can save, uh, recipes. It's like a GPT, so prompt you save that. You can run every time on the meeting. so, and you can run on them on the meeting, on the folder or on everything. So for instance, one fun thing I've done for my clients at the end of the year, I've created a recipe, which is, uh, like the Spotify wrap is the year end review.

[00:36:31] Yeah.

[00:36:31] I created a prompt that says, Hey, look through all the meetings with this person. identify all the strategic topics we covered. The progress they made, the blockers they had, uh, the topics were addressed most and the topics that you recommend we tackle the most next year. And then I also did the Windsical Windsical stuff, identify if they were a famous song, it was their spirit animal was they go to catchphrase, whereas the, the word they should be using less.

[00:36:56] And I created just very simple, word documents, or at least Google Docs that I put in an email and send them as a end of year.

[00:37:04] I love it. I, you know what, I want one, one more step. If it went straight from your program, if you will, to an email draft, that would be killer, right?

[00:37:13] Uh, yes, but I, I did review everything 'cause I didn't want, I wanted to make sure nothing was gonna be offensive, of course, offensive to them,

[00:37:19] um,

[00:37:20] especially the spirit animal and the, the, the song I wanted to make sure it was okay.

[00:37:24] Oh, this is huge. Also for all of you who've got sales team managing strategic account. If you can find something like this, that if you record every call could do a quarterly recap of the progress they're making, uh, there's ways you can really instrument something really engaging for customers.

[00:37:41] And so a lot of companies have systems like Gong, which are getting, you know, where they're recording their sales calls and then they're running. Those are starting place for some amazing, sophisticated things, uh, looking at win win-loss rates and what happened in those calls. helping provide tutorials for the sales salespeople who were off message and not, not doing it well.

[00:38:03] Customer insights, competitive alerts. There's a lot of cool things that happen there that you know if, if you're really putting a brain on top of gong it, it can be a huge asset. I'm wondering, is that the same thing that in theory, you could do because you could take all those conversations, upload those conversations, and start to do trends analysis?

[00:38:26] Yeah, absolutely. So you can then create these prompts, either one off, or you save them and you run them to analyze something, look for trends or what, what, what I really, really think is great for marketing teams to, to think about is how can we turn this into a high value added value asset for the customer?

[00:38:45] Something that helps them learn, think about further, just, gives them extra insights, et cetera, and feel special.

[00:38:53] Well, and and I think that's a really good transition to what I think is, so a lot of AI has been a tool in search of a problem or a strategy. And if we say that the strategy of the organization is put the customer at the center, and how do we use AI to help the customer, that to me is a very liberating approach.

[00:39:18] Now we're not looking at just efficiency because efficiency's good 'cause that allows us to maybe lower our costs or service faster. But we're thinking about tools that we can, the recap that you described, and I'm curious, in your conversations with CMOs and some of the things you're doing, are there some AI really cool things that some of your, your customers are doing to sort of in the name of customer centricity?

[00:39:44] Yeah. So for instance, there, there's one I work with who, who created in cybersecurity a fantastic maturity model for the security cybersecurity discipline. They're, they're encouraging and, and empowering and we, I initially, I built them a quick prototype. So they put the maturity context, all the, the white paper into, uh, cloud.

[00:40:05] And I told her, can you vibe, code, uh, the equivalent of the survey? So that people can quickly insert their surveys, see what stage they're at, and as a result, automatically, uh, speeds out recommendations for how to get to the stage they're at to the next. Uh, they thought it was pretty cool and they created that which is now on their website, uh, because I couldn't, uh, instrument in a way that would scale and

[00:40:26] but this is value add, which turns something that's, you know, when you read maturity models is not super engaging. No. All of sudden becomes. This, uh, survey, this interactive tool that they can take. And by the way, as they complete this, you, you further qualify them like crazy for follow up with sales or campaigns.

[00:40:43] And that's the type of value add you can create. Just like our, the RI calculators will build and I can also be a little more engaging if you vibe code them. you can also, structure. a lot of information about it and organize them by vertical and try to recommend and find the right insights to give back to a given customer in that vertical where you take, these are the solutions that are right for you.

[00:41:04] The right RI calculators, the right case studies, the right quotes, you can package that up, then give it to sales. They review and, they send over if it fails, right.

[00:41:13] small thing, but, and quickly, so you've run hackathons and training sessions with teams. Yeah. What are you seeing that's working that, other than the CMO participating actively, to accelerate AI adoption?

[00:41:25] Yeah. The, the first thing for these hackathons is, um, make sure that you, there's a team that prepares them by giving access to the right tools first, making sure that it's set up. So whatever you use, make sure that they're available that day. To, start with a few presentations before the actual hackathon where someone, the exec probably gives the mandate and the encouragement.

[00:41:46] Then I recommend you have two to four more presentations where people show their workflows, like, like I was talking about earlier. They inspire others. Like, this is what I built. This is how we did it and pretty specifically show them how they did that. Then maybe one or two people who know a solution very well could be it who set up your, your Glen account or whatever tool they do a little demo and enablement of that.

[00:42:07] Then you want to put teams together where you've got folks who are advanced and folks who are not as advanced together so they can inspire each other, and. At the end, you get them to present, there's a jury, you give gifts, make sure the execs are there, showing the results. 'cause every time I've been, I've seen those, there's always an exec that says, I don't know if you win or not, but I want to fund this.

[00:42:28] And, it just, it gets the team energized. But my advice is then when you've done that, just don't stop and wait for a whole quarter before doing this. Keep Making times and blocking times. It can be a Friday afternoon, every other week, whatever, for teams to continue like that and then create small powers that are going, continue to work together, on this project and give them budget.

[00:42:46] This is where AI trainers can also help and, and Right. And make sure that they give guidance and be there during the hackathon.

[00:42:51] I wanna make sure that in this process that there's a strategy and a goal that we're, there's outcomes that we're trying to do. There's customers that we're trying to serve.

[00:43:00] There's problems that we're trying to solve that either internally or externally, that ladder up to something big. 'cause all of us could build cool things. Right. Yeah. But they may or not move this so that there's an infinite number of tools. There's an infinite number of things that we could do. And by and by the way, I do think that's the downside of of this is there are a lot of folks that are spending 18 hours a day.

[00:43:23] 'cause it's just so cool to make stuff.

[00:43:25] Great point Drew. And, uh, at Twilio we're doing, um, we have a team that was specializing in organizing in, helping very large strategic accounts do hackathons. And the way to your point we would prepare them is looking through S ones and filings and working with the execs to identify what are the top priorities that you, like your team, that your team can pick up on, on this project.

[00:43:45] The hackathon strategic priorities and things that ideally they would take on. The teams don't have to take them on, they can take something else, but at least there's the suggestions of big, uh, strategic problems to take on.

[00:43:56] Okay. Um, Isabelle Pius, would you like to ask another question?

[00:44:00] I find that personal experiences, we all, you know, I come out of a lot of these AI conversations and it's. The amount of use cases and all the things that we could be doing is overwhelming, right? Like I think it's limitless. And what's missing from the conversation consistently is the strategy is the manifesto.

[00:44:16] As a company, why are we using ai? Right? And so thank you. It's like, so it's really pushing what are the core strategic priorities and then just anchoring everything around that. Otherwise, it just becomes just too much really.

[00:44:28] Yeah,

[00:44:30] that, that's a very good way to do it. Now there's also things that don't look like a strategic priority, but are amazingly manual routine and weighing your team down.

[00:44:39] So it's important to let also surface, from bottom up and let them list that, and previ given them the time to, to work on those and be, because even for myself, there's stuff that I realize I always put off to the end of the day, the end of the week, or never get to. And at some point I was like, well, let's me spend two to three hours, maybe more.

[00:44:58] Actually making sure automate and making sure I automate it. And then once I'm done, like, oh my God, my life is so much better. Now I can spend time on more strategic things.

[00:45:05] probably probably why you want to come in and prioritize.

[00:45:08] Yeah. I mean, I do think there are lots of, you know, a stage, you know, stage one is everybody's using it for content, but even within content there are workflows that are really messy that could be greatly improved.

[00:45:19] And then there are, so I do think that an audit of workflows is in order. You're right, because those are some easy things. Uh, and one example, uh, that we talked about at another, at a strategy lab was that your funnel right now is probably incredibly inefficient. And if you could a, use AI to somehow or other make it more efficient, that's a worthy thing.

[00:45:43] And by that I mean you have a hundred thousand leads coming in and that becomes ambi, qls, and it, and finally you get down to 20. You know, close one or whatever. That's all huge amount of waste. And so if you could figure out a way to use AI to drive efficiency there, that would be a very sophisticated use case.

[00:46:01] Right. Have you seen anybody do that?

[00:46:03] Uh, yes actually. The, the AI hackathon that was a judge, that, there's a team that actually did exactly that, they, they gave, uh, AI access to their, uh, Salesforce and identify where, identify where we have opportunities. To increase convers rate from one step to the next, what to do, et cetera.

[00:46:21] and that, that was one, one of the ones where the CEO raised their hands. Like, if you don't win, I'll fund it. And you should have seen the team was so energized.

[00:46:30] Yeah. No, it's, that's great.

[00:46:33] Coming up with the idea and seeing it was possible.

[00:46:35] No. And, and that when, when you get to those kinds of things.

[00:46:39] you're really helping the business move forward and I think to sort of Isabelle's point as the leader, again, your job is to allocate resources and make sure that there's focus on the big things that matter and obviously efficiency of pipeline. And I was sort of, it's no different than having a better defined ICP.

[00:47:03] If you have a better defined ICP, then you can screen out all the ones that aren't in it, right? And you are better and better, better at closing those things. So focus matters, but in order to focus, one of the things that's clear to me in this conversation is you need to have some sense of the art of the possible.

[00:47:21] Yeah. Yes.

[00:47:22] And I think that's really hard right now because it's moving so damn fast.

[00:47:27] It is hard. Now, there's some resources that are really good at recapping some stuff. So for instance, I don't know how many of you watch or listen to Marketing Against the Grain by, um, the SVP and the CMO at HubSpot, but it's inspiring to see the amount of time they spent themselves and.

[00:47:43] Playing with the tools, learning it, and then sharing that back to others. And when you do something like that, of course they're, they spend a lot of time on it, but it's essential for their business. 'cause they're always been a content business, a thought leadership business, also a tech business. it really sends the signal to the rest of the team, like, if my CMO can find the time to do that, maybe I should too.

[00:48:02] and they understand the area possible. So Marketing Against The Grain is a good resource. There's other ones that you could find or I can recommend. That are not so long to consume. And what I've noticed, the more you understand the art of possible, the more you can put two and two together and like have creative ideas or allow your team to have creative ideas.

[00:48:17] 'cause you realize, oh yeah, I can take this match with this and we can build that.

[00:48:21] and it's really fun

[00:48:22] and that's one of the superpowers of most CMOs is they can connect dots that others can't.

[00:48:28] this should create those opportunities. It just doesn't feel like it always, uh, right now at least.

[00:48:33] Alright, we are running outta time and I, there's so much more that we need to cover, so obviously we're gonna need to do this again. let's give, uh, the folks here two key takeaways that you want every CMO listening to Remember, when it comes to becoming an AI augmented CMO

[00:48:52] Again, block the time at the same time other people on your team.

[00:48:55] So it feels easier to do that and identify your key champions and ideally champions plus angels. So those are gonna be okay spending time with you and others showing you how to do stuff and helping you be your partner. and be okay to be vulnerable. Just say it's okay to when you don't know, 'cause no one knows it's moving too fast.

[00:49:14] I bet you even the people who work at the AI labs can't even keep up with their latest tools.

[00:49:18] so it's okay to see, I don't know. And this is where you need to just, have someone else, help you and tell them I need help.

[00:49:24] I love it. Alright, well Fran Devour, thank you for joining us. Where can people find you

[00:49:29] Seems like they can find me on LinkedIn, of course. And then I also, uh, publish a newsletter that you can also find on LinkedIn once every other week I share some of some of the things that I learned and I see from others. Um, and um, yeah, that'd be great to go.

[00:49:43] Awesome.

[00:49:44] Awesome. Alright, well I do hope that you've put, um, October 22nd and 23rd on your calendar.

[00:49:50] We'll be in Palo Alto for the CMO Super huddle.

 

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

Show Credits

Renegade Marketers Unite is written and directed by Drew Neisser. Hey, that's me! This show is produced by Melissa Caffrey, Laura Parkyn, and Ishar Cuevas. The music is by the amazing Burns Twins and the intro Voice Over is Linda Cornelius. To find the transcripts of all episodes, suggest future guests, or learn more about B2B branding, CMO Huddles, or my CMO coaching service, check out renegade.com. I'm your host, Drew Neisser. And until next time, keep those Renegade thinking caps on and strong!