IBM's AI Marketing Journey from Experimentation to Intentional Execution with Craig Mills

Shaheen Hoda: He set it up on a remote computer and he's logging from his phone. I haven't seen anything as addictive as Diablo.

Shaheen Hoda: Right.

Shaheen Hoda: The video game Diablo did, uh, and I'm guessing her husband played a lot of it.

Craig Mills: I played a lot of that.

Shaheen Hoda: No, you did.

Shaheen Hoda: Welcome to APAC's B2B Growth Podcast by X Growth. I'm Shaheen Hoda, and um, on this show we're going to be unpacking trends, busting myths, and delivering insights you can put to work today to, to drive growth in apac.

Shaheen Hoda: There is so much noise around AI

Shaheen Hoda: right now, so much theory, so many

Shaheen Hoda: thought leaders on LinkedIn promoting all their agents that do everything in their life

Shaheen Hoda: and company, and so few honest accounts

Shaheen Hoda: of what, uh, it actually takes to make it work inside a real organization.

Shaheen Hoda: My guest today, Craig Mills, is the APAC head of Demand generation and AI Marketing Transformation Lead at IBM, and he

Shaheen Hoda: has that honest account.

Shaheen Hoda: We're going to talk about lockdowns of AI use in the organization to false starts, to promoting adoption internally and all the hard lessons that followed. You don't want to miss this episode.

Shaheen Hoda: Let's dive in. Craig, thanks so much for coming on the podcast.

Craig Mills: I'm really happy to be here.

Shaheen Hoda: I'm very excited and um, I'm really excited about the topic, the IBM's agentic mission. I am very excited to have a chat about this. I thank you. Bring a lens that not a lot of people get exposed to. So, uh, um, very, very keen to hear your thoughts and experience on, on this mission that IBM has been on. Obviously when we're, when we're looking at AI, everybody's talking about it. Whether you go to a conference or you catch up for a beer with friends, you're like, oh, you can't believe what I've done with this. I'll, uh, put a couple of things in. And so I'm really interested in, you know, IBM obviously sells enterprise AI into the market. I also know from our previous conversations that you, uh, were client zero. Tell me a little bit about that.

Craig Mills: We've been in AI for a long time. Like before the current generative blitz like predictive analytics has been around forever. So we've been at the forefront of AI from the very beginning. No question on that. But it really got a tailwind, of course, at the GPT moment. And what that did is change the imperative within the company. And so from the very start it was, uh, go, uh, and go hard. Now that in itself is not enough. It's how you then apply it and it's one of those things that I really take a look at the different functions within a business and something we're trying to bring to our go to market strategy, but also then how we apply that internally, the differences between e.g. hR, IT marketing, M sales, supply chain procurement, et cetera. What we found internally is that HR was the function that appeared and seemed to get most traction certainly within the corporation really and how it moved fastest. And so uh, we noticed this as employees. That's where it became very, very, very apparent and very real for us. So while the whole world was out there talking GPT and what I did for my travel and how I did uh, this latest research and the recipes I got, we were experiencing that for real as employees internally within the enterprise space. So our HR systems changed substantially over the last five years. And it goes back pre GPT by the way. So that's why when GPT hit it kind of went and felt certainly to me it was like okay, what's the big deal? Because we'd already been on that trajectory. But. And this is the but in it uh, is that we'd found uh, that we were going too slow. It was kind of like an incremental build and where we were going well and we were doing it at scale and we were going methodically through it. The world shifted up a gear. And that in some ways was the biggest challenge for us to come to terms with. I think we still are and the entire world is still dealing with it, which is the intensity that is now here. So there was always a trajectory of AI. It's always been there. It's the rate, the pace, the intensity and now most importantly the returns on that effort that are really under the spotlight. And when I came in here I'm um, feeling upbeat at the moment because a whole lot of great things happened this week on a personal and a professional level. But one of those things on a professional level is that we made a bunch of decisions internally to then ramp it up in other gear. And what I'm really excited about is how we've come from a, ah, I would suggest a very traditional methodical enterprise IBM approach. And that's what we've always done. We've been probably slower in most organizations, but more methodical. And that gives us an element of trust. And the baseline of trust of IBM is paramount to now that uh, trust plus acceleration, like we want to use that baseline to then go super fast and to go fast at enterprise size is bloody hard. Right?

Shaheen Hoda: That is, I want to talk about that. Right. Um, but I'm also curious to know, I don't know what you can share. When you say ramping it up, what does that mean?

Craig Mills: That means, uh, for example, taking more risk taking. So again, we're a risk averse organization, have been forever. It means deliberately taking more risk. And the processes of I better check with this person, I better check with that person. I better check with this person. Those guardrails are being taken down now again, both at the enterprise level across all functions. And then within marketing, we want to be very thoughtful about how we take those guardrails. I won't say down, but loosen them to a degree. Because if we don't, we get beaten in the marketplace. And getting beaten in the marketplace at the moment is not an option. I mean, you've heard this sass apocalypse story going on at the moment that's affecting us as a business. Like we're a software player now. We haven't been as hit as hard as some, as the others. But our stock price is down. That's public information. I can talk about that. But it's our reaction to that which I'm seeing as a really positive outcome. You know, senior leadership discussions we've been having during the week on how do we go faster, how do we take what we've done, which has been really good to date, and then just quicker, faster, better. It's just quicker, faster, better. That's the world we're in.

Shaheen Hoda: I want to come back to this, but what I also love to hear from your side is a little bit of that journey right when you. Again, it was being implemented in the organization from early on. But Chat GPT is really OpenAI is really. So Chat GPT was probably a little bit of a catalyst. And from memory there was a little bit of an arc from what you told me at uh, IBM, where at the beginning it was a little bit different than in the middle than later. Can you tell us a little bit about kind of what happened in the organization in terms of the adoption from the early days? Right. Like um, in 2022 when the kind of wave start to uh, rise. What happened in the organization until kind of what did that arc look like?

Craig Mills: Yeah, the arc's been very interesting and I'm um, writing about this as a sort of a historical reflection because I'm a big believer in history, by the way. I love history and politics and geopolitics and all that sort of work. I went back and wrote down all my thoughts and then started constructing them into almost a, uh, history of what happened. And it goes sort of along this timeline is that like I said, up until about 2022 it was business as usual, which was predictive analytics. That's what it was. The GPT Moment hit late 2022 and it was like, well we at the time. No AI time.

Shaheen Hoda: Lockdown. Lockdown, yep.

Craig Mills: Okay.

Shaheen Hoda: I think a lot of organizations did

Craig Mills: and it was just sit back and say now of course you can't stop IBM people from tinkering in the background. Right? So behind the scenes there's a lot of stuff going on. Right. So it was this unwritten rule of don't do it, but then very carefully do it for your own personal gain and personal knowledge, but be careful, you can't use that in market. And that was okay. We then responded in May 2023 with WatsonX, which was IBM's response to the GPT moment, which was we needed to build a product platform. And we announced that at that time that really did change it for us. And it didn't take long mind you, because that was only three or four months later an entire new platform. But again within marketing we couldn't really use it. There was not a lot to be able to be done. That again changed in around 2024 where we then enabled the entire organization on the what's next platform. We had an internal, always called what's next challenge. Enterprise wide, 300,000 plus employees using AI. But again, to be completely honest with you, uh, we couldn't use it from marketing perspective. And it wasn't that we weren't allowed to. It was, it just had no practical application. It was very IT oriented and we all sort of went, is that it? And then a couple of other tools came online and what began to get enterprise grade came along. Most particularly we had an internal tool called Consulting Advantage. The consulting group developed it and it was based on all the external LLMs, but they were cleaned and put internally so we could start using it. That changed the game internally for us. Then we got Copilot, then we got Box AI and these enterprise suite of tools started to come along. And from that moment then all hell broke loose. It was like just flowering of ideas and innovation and people trying all sorts of things. Now we harnessed that through the latter part of 2024 into 2025 with what we call the AI for marketing group within the marketing division. And we brought a whole group along with us to understand the practical implications using AI for marketing. Some of it was sanctioned, some of it was freeform. And there was, I would suggest a really bright flowering. It was like the springtime of ideas. It sort of went nuts. It was great and then it died off because without the enterprise grade tools that were truly sanctioned and able to be used with guardrails, we couldn't apply and what we needed to add into the marketplace at scale. And that happened all through 2025 and now 2026 has come and it was almost like an ebbing towards the end of 2025. It's like we'd run out of the informal implications of what we're doing come 2026. Now we've got to re engage ourselves and get much tighter on real AI for business value. And that is the mission we're now on. And that's why I was saying we've had a bunch of things happen earlier in the week and over the last few weeks which has really put a rocket under that. So I'm really excited about how we're going to do that.

Shaheen Hoda: So how have you changed your approach to the use of AI? Let's say, you know, in marketing, you know, it's, it's, it's your domain. How have you, I don't know, incentivize teams or kind of mandated if you put any kind of restrictions in place or restrictions is not the right word but uh, elements that would drive a certain behavior.

Craig Mills: Yeah. And that's talking through that arc. And the really important piece is probably, let's say 2024 was. And 20 was learning. It was just trying to figure out which way's up and then how to even conceptualize some of these things. 2025 was all experimentation. We ran internal challenges. We had 300 plus people work on our Adobe. We're in Adobe shop on the Adobe suite of tools. Firefly, our internal consulting advantage copilot box, the whole works on how to just ideate. We kept some of that only for internal use but then under certain conditions we could use that externally as well. And that was brilliant because we got this, like I said, this broad springtime of flowering of ideas. But I'll be completely honest, again, some of those were just an absolute waste of time. We wasted a lot of time, not so much money. But it was more time on things that just didn't really move the needle in terms of our marketing. Come to 2026. It's about getting intentional and very deliberate on the use cases. The use cases that we believe will drive either two things. Really simple growth. So top line revenue growth of the company or productivity cost takeout. Now the idea is if we can quantify those very specific use cases, decide which ones are the best ones that are either going to drive that growth outcome or that productivity dividend. Then we need to rally the organization around those. So it's kind of like a cleaning out process of the portfolio. I'm looking at it like a portfolio now, which is you might have a thousand things in the portfolio which we do. Right. Every different micro use case that someone's developed of their own initiative. How do you then clean out the 90% of that which is marginal to low value? In fact sometimes there's negative value. It actually strips productivity out of the organization into just those use cases that are going to move the needle for the company.

Shaheen Hoda: Are you able to give an example of huh.

Craig Mills: That yeah. Email. Email for me is one of those use cases that is very, very easy to apply. And this was part of the experimentation process and I've written about this as well where we've uncovered that email is one of the most effective but probably least utilized channels within IBM for a number of historical rounds. Yeah. Now coming into 2026, email is my opinion one of those strategic weapons. We can really go with email, email

Shaheen Hoda: and tell me why you break it down for me.

Craig Mills: And this is the point. Right? The point is when I see a lot of the AI thought leadership that comes out and all of the ideas, they're big, complicated, uh, multi channel, multi this, multi that, dimensions and agentic this that. What I mean by that is there's a lot of big theoretical stuff that people like to talk about and use buzzwords to try and pretend like they know what they're doing and they're talking about. What I tend to find is the practical application in certain areas is the key. Like what can you actually do with this? Now at an enterprise grade my opinion is email is one of those.

Shaheen Hoda: And that's like. Are ah, you talking about like email email marketing? Yeah. Okay. That's where email marketing, the communication between I don' as SDR or kind of marketing nurture or that. That's the air. Okay. That's what you talk.

Craig Mills: Precisely. So we generally divide our email internally into two types. There's inbound nurture email and then there's outbound email. And the outbound email gets done sometimes by our uh, marketing function and of course also by our sales function. Now it is a complicated and fragmented system. My opinion is that AI is one of those tools that can fundamentally shift the way we perform on our email. And we've proven it in small scale, in pilots and in small campaigns. Next step is to take that broad strain and actually systematize that. Now we have global email organization and they're a fantastic organization. They're working on this from a corporate.

Shaheen Hoda: There's a global email organization at IBM. Oh yeah, that's one of the things that only IBM would have like large companies that have the global email organization.

Craig Mills: So we have, we have what is our automation practice and that is email. Right. And it's all marketo based or Adobe Shop. Um, but it also then has to integrate with what we're doing out of Salesforce, Salesforce Shop as well, um, as well as other tools, Salesloft for example. So we've got all these different divisions doing different things and we're trying to bring that all together in a common framework. Now that's going to happen at a corporate level. Our job, because I'm in the geography team, my job is to do stuff that drives revenue in the markets out in the field. I need to connect the dots between that broad IBM corporate initiatives on email optimization. Automation and AI of course is a key component. But how do you do that in the field if you're in India, Australia, Korea, China, Vietnam, Indonesia and how do you do that across what is a portfolio of uh, over 100 products? And then we've got the services division and we work with everybody from the largest governments, airlines, banks in every different country through to individuals, people who buy just one product. Like our breadth is extreme. Trying to manage a multi dimensional marketing mix across that framework is really hard. So narrow it down. And email is one of those. Ah, now that's a huge piece of

Shaheen Hoda: work by the way I would imagine. I mean at your scale, right? Like at our scale if we're like a company of like 30 people, sure that's, that would be a lot simpler to roll out. But when you are, when you add a couple of zeros to that and become 30,000 or 300,000. Yeah, all that, that is the simplest things become. It's like the whole theory behind building a car. And it's like the larger the vehicle becomes, the complexity of the engine exponentially goes up. Um, and it's the same thing in a human organization.

Craig Mills: It's exactly the same thing. And so complexity of email at enterprise grade is really tough. And then this is where there's the knock on effects too because then you start to look at email as it relates to product trials and demos. So we've got to figure out what the sign up scheme, the nurture scheme for our products and demos are so that then Comes into email as well. Then you look at event invitations when we do a lot of events. How does that work? There are so many knock on effects when you think through what is email now again putting boundaries around what it is we're going to do and how we're going is the key. So one which just kicked off this week. Just like real life.

Shaheen Hoda: Let's go.

Craig Mills: We know that we get huge numbers and this is mostly in India at the moment, but we get huge numbers of responses coming in this quarter alone, 15, 20,000 responses and responses are to responses to our.

Shaheen Hoda: We're talking like tickets.

Craig Mills: No, sorry, not tickets. Response and marketing responses.

Shaheen Hoda: Oh I see.

Craig Mills: So people who do the trial download a white paper.

Shaheen Hoda: I see.

Craig Mills: Come to a webinar, not an mql. That's different. Those are just the raw responses. We just get literally tens of thousands of those things. This is just for our agentic AI campaign that we've been running. So 10,000 plus responses in a quarter and that's filtering out the students, filtering out our business partners, filtering out the individuals, these enterprise responses. We're now trying to work out how to take each of those. And the big learning for our uh, agentic AI campaign is that at last, as I was saying before, the industry has now moved from experimentation just like we have internally as well into practical application, return on investment. And so we were campaigning on general statements last year and again most of earlier this year we've made the very deliberate decision to go very, very deep on specific use cases. Again we're bringing it back to use case. Use case for me is the building block now of AI both in our go to market as well as our internal development. And so we're now building our uh, response to managing all those responses and the nurture strengths for those responses into the most relevant and useful use cases in the agentix space. And what we think, we can't prove this at the moment but we think it's supply chain, hr, procurement and finance. They're the big ones because they're the repeatable process driven functions, easier to quantify, easy to draw out as a business process and then turn into an agentic process. Notice that sales and marketing is not on that list.

Shaheen Hoda: Yeah, uh, or software developer.

Craig Mills: Software Software development. Let me bring that one up. That's, that's a very different category for us. I'll get to that in a moment because that gets back to my SaaS apocalypse that I was talking about. So business processes excluding software development because that's in a very special category all of its own. So then with email, this is where we're staying at to be relevant to those audiences. And it's all about relevance, giving it back to Marketing 101. Right. We often forget the Marketing 101 lessons. Right content, right place, right time, right audience, right. We now have micro segments of each of those roles, disciplines, industries, in some cases territories. Like you take North India versus South India, you take Western Australia versus Eastern Seaboard. Very, very different approaches. We are now going to segment out each of those little groups into then micro segments and then using our AI capabilities, micro target those and most importantly, measure them. Um, now we did that in a few pilots last year. Amazing results. And this is what somewhat frustrates me as well.

Shaheen Hoda: I love, I want to hear this because one of the things I wrote down is what is frustrating you?

Craig Mills: Yeah, frustrating me is uncovering these moments of gold and then not applying them at scale. And so we did this last year M. Like we did this before and then we got distracted by the other stuff and it's like, no, no, this is where the gains are. This is where the easy money is. So we're going back to get the easy money before we try and go off and do something fantastic and groundbreaking and change the world. Now that's good because sometimes you have to do that. Sometimes you just got to take the easy gains where they are. So that's email. So we can go back and do email properly, do it at scale, institutionalize it and move on to the next thing.

Shaheen Hoda: This concept of micro audiences and micro segments is a very fascinating one. One that I've been thinking about, um, the past couple of months. Like we do a lot of ABM and targeted programs that is very kind of curated, but for the longest time you couldn't scale that. Like that was for your, uh, depending on the type program you're running for your first five biggest clients or the most, you know, the 20 most valuable accounts or you know, maybe 100. But this idea of micro segmentation and micro audiences where you're like, I'm going to take this pocket of, of people that have so much in common that previously the ROI was not there for me to kind of do that level of personalization for and then creating hundreds of those and, and they go in that market. It's a very, very fascinating concept. And, and maybe uh, my follow up question to you on that is, are you looking at this from an enterprise lens? You looking at it from. Tell me a little bit about like how you're kind of thinking about this from the market you talked about geographical, but does it go into departments of the organization? Uh, tell me a little bit about how, how deep the rabbit hole goes.

Craig Mills: The rabbit hole goes very deep and that's what makes there has to be a point of diminishing returns. Like there's a point where if you take it too far there's no more juice to squeeze. So therefore is there, is there though? I don't know. This is the thing and this is where paradigm shifts are uh, just being made in every dimension in business at the moment. Whereas the old assumption was what I just said which is you get to a certain point where there's just three

Shaheen Hoda: years ago, yeah, 100%, maybe that's not true anymore.

Craig Mills: Maybe using the AI systems that uh, will self learn, self manage, self. Correct. They will then squeeze quite literally to the point where there is no juice. Like it's done, it's literally dry. It's not just juiced out, it's juiced out. I believe we're going to start on email to squeeze as hard as we can because I see it as the most repeatable process in marketing. Now I want to get in touch on that if that's okay for a moment because that's the very narrow view. You then blow that back up again. Which is again the thing that I've observed from an industry perspective and certainly from a function and discipline within marketing, certainly at IBM M is that again functions like hr, supply chain procurement, so forth, they're very repeatable. Marketing has that element of creativity in there is one of the few functions that you've got to find that perfect blend of workflow which is process driven productivity enabled output multiplied by uh, the great unknown which is creativity, individual creativity and organizational creativity. And it's the magic source, right? It's what makes some companies amazing and others average. Marketing is that blend where you've got to get the two right. And I actually think it's one of the reasons why as a function we've probably struggled to apply AI at scale. Because while it's easy to quantify and draw out a business process, it's much harder to quantify and draw out creativity. M that's what was traditionally seen as a human element. Now again, old assumptions. When does that shift? How does that shift? We don't actually know that yet. But what I do know is that the more process driven parts of a marketing function or where to start, email's one of those creativity. I'm not sure where that begins. So I'm going to Leave that one alone for a little while.

Shaheen Hoda: You got a number two after email.

Craig Mills: Um, media, I think paid media is one of those ones. And our paid way again, we have a global paid media group who does that. I did a project in the very early days where we played around with paid media.

Shaheen Hoda: Which part of it? As in like placement or just like creative? Oh, creative. Creating the creative. Creative.

Craig Mills: I see. Yeah, yeah. And again, what we did though is not really work. The creative element. This is what formed this thought in my head, sort of reflecting on the experiences. What we did is we enhanced the process inside of paid media. So when we put together a paid media campaign, we'd obviously need to get, you know, the simple executions, banner ads, things like that, particularly for display. What we found is when we did, uh, these new creative tools. No new at the time. Now it's sort of old. But what we found is when we mapped out the process, particularly through the different committees that need to be engaged. It was up to a six week process to get things from ideation into market. Six weeks. And there were so many different groups involved. What we realized is if you take some of those groups out of the loop for a start and then you replace them with an AI, uh, interface and one of those was legal. And I don't be careful how I say this, because we didn't replace legal with AI. Do you know what I'm saying? We love legal people. What we did is we simply rate it because we knew what the legal rules were, they were codified, right?

Shaheen Hoda: Yeah.

Craig Mills: So what we did is just build a simple reference point which then put the ad through their legal interface and it just rated the ads as high, medium or low risk. And anything that was low risk went straight through. And the legal team were happy with that. M They're like, yeah, that's just complex straight away. And we did that and a few other process steps.

Shaheen Hoda: Love it.

Craig Mills: Uh, and suddenly we reduced from six weeks to a week.

Shaheen Hoda: Wow.

Craig Mills: Now, because we were able to do that, we then had to be able to do multivariates. So instead of having five, because that's all we would send to legal because it would take forever for them to go, there are five, we could now send 30 and we obviously have to get the top 10 still done, but the other 20 straight through. Now those variants then gave us better, uh, options in the market. We did A, B, C, D, E, F testing and of course we moved to the most optimized ones which then drove the performance up. So on two levels, that's my point. About productivity and growth. A, we got productive, and B, we got much, much better results. More bang for our buck. And so we're able to drive better results for the company.

Shaheen Hoda: Craig, I mean, I think this is great what you're describing, right? And, um, and I think it's a, it's an area that a lot of organizations want to get to. And what by that, I mean a situation where, hey, we've done the experimentation, we've seen this works, now we're going to scale it and we've scaled it, we've seen the positive impact. I do feel that that's not necessarily what's happening in a lot of organizations. And the scaling part is the part that is proving not so easy. I mean, you know, especially for a lot of our clients where. And some of them are in the news with what they're doing with A and all that stuff. And. But it's internally, it's not as smooth as one thinks. For, like, sees from outside of like, oh, this is a nirvana. This is like, oh my God, they just put it in here and just go. And the agents jumps on it and puts it together and then it goes out and then the client sees it and converts into a customer and it comes out and pipeline and targets met, all that stuff, right? What is, what is your take? What are you seeing, right, for, um, the actual implementation of these things and really come into life insider companies. Because there's a lot of promise, right?

Craig Mills: But yeah, how do I say this the right way? There is so much BS out there. It is so hard to do. Anybody who says it's easy is not doing it. They're kidding themselves. Actually doing AI at enterprise scale across multiple industries, multiple products, multiple countries, multiple even domains within marketing, event media, email, et cetera, that sort of stuff, it is so difficult to do. It's so difficult on so many different fronts. So my big message, I suppose, is it is grind. It is hard work. But, uh, and I'm naturally positive it is so much fun, uh, because I get to play with this stuff almost every day. And it's intellectually really, really challenging. And that's why I say it's hard, but it's fun. I like to think of it as the hard gains you get when you do really deep exercise as well. Like when you go for a run and you've really pushed yourself or you ride, flat chat or you swim, whatever your thing is, if you consistently get to that point where you're enjoying it because it's the challenge. That's what I'M finding with the AI work I'm doing at the m moment, I almost come home every night. No, I'm still going in the nighttime as well. Then wake up the next morning like, oh, uh, right. What's next?

Shaheen Hoda: My wife is not happy. Yeah.

Craig Mills: But it's really, really invigorating because it is just such an intellectual challenge, intellectually, to keep up. Just, uh, every day there's something I'm like, oh, my God, how am I going to ingest that? Think about it, understand it, and then apply it, and then the application, then to actually apply it with discipline. And that's my point before about email. I sort of feel a little annoyed at myself that I let that one go for a little while and now it's like, no, get back to it. That's got real value. Like, let's go into that.

Shaheen Hoda: I want to come back to what makes it hard just to make this happen in enterprise. Right. But when you kind of said it's addictive, it's very addictive. And I was in a group chat and one of the members of the WhatsApp group, like, her husband, is on this all the time, right? And she's like, he set it up on a remote computer. He's logging in from his phone. She's like, I haven't seen anything as addictive as Diablo. Um, right. The video game Diablo that I don't know how many years ago came up and I'm guessing her husband played a lot of it. Um, but she's like, I haven't seen anything as addictive as Diablo until. Until, uh, this thing came.

Craig Mills: I played a lot of that.

Shaheen Hoda: No, you did.

Shaheen Hoda: I love it, I love it. Um, uh, but tell me a little bit about what makes this so hard in enterprise.

Craig Mills: What makes it hard is the number of choices we need to make at enterprise across different market segments. And again, maybe this is just an IBM statement because we're across so many different segments. We truly do span, like I said before, huge, biggest enterprises, governments, telcos, etcetera through to individuals. So that's tough for us. But what I think makes it the hardest is the interconnectivity that's required to do AI at scaling. What I mean by interconnectivity? The interconnectivity of people, process and tools. So, for example, the tool sets, right? And I don't want to get into good, bad or otherwise of tool sets, but tools have to work together to do AI and agentic AI at scale, and that is not easy to do. The technical implementation is really tough. And I'm not a technologist, although I'm dangerous enough. But even I uh, tried to do that internally ourselves, right. This week I've, you know, one of these addictive moments.

Shaheen Hoda: I had a few, the guardrails came

Craig Mills: up, seriously, the kids were out and I was, you know, I was blood feeling there and I had to play with this, right? So I went and I played with, I uh, won't go through which ones. But I went and tried to connect two tools that should have worked and they didn't. And I'm not an idiot, right? I'm like, oh damn it, I got to do, I've got to spend more time on this. And I spent two to three hours just following the rules and it still didn't work. And like uh, so that, that's one thing. What should have been enterprise grade tools, well known tools, just didn't work as they should have. I then wanted to take that and present it to the team because then the second piece is the people, right? People have different levels of engagement and ability to then take this to the next level as well. And what we found internally is that there's roughly, I would call them 5 to 10% of the population which are uh, probably a little bit like me, just in the rabbit hole, willing to experiment the advanced users. And at IBM M we've got a pretty good population of then very technically literate people who will apply at scale and I'd call that another 70%. And then we still do have those people who are just not that interested. We have to empower that bulk, that middle group and then back to the tools. I found it difficult to do this thing so therefore I couldn't empower those people, uh, which was frustrating for me. So then I went and found a different use case which was a bit easier. It was a single tool use and I used that instead.

Shaheen Hoda: Yeah, right.

Craig Mills: Uh, and then you get to the process side of things which is okay, so you've got a tool or a couple of tools that work together. You've then got people who are using those tools, but are they doing in a standard way that you can then quantify and then extract real business value from? And again it's that cumulative effect of the tools, the people and the processes and they also work in different permutations, combinations. Sometimes it starts with the process that's the people, that's the tools, vice versa. Landing on how to operate that system, uh, is complicated and it's hard. That to me is the real difficulty. But what you said Before, I think is the key is that when you find something that's, uh, so complex, the only way you can deal with it is in a simple way. And I bring it back now to that simple way is the most valuable use case. So for me, bring it back to email. Now we'll pick probably three use cases in our division. Tend to really narrow on email's one already. I'm going to go and decide with our leadership team what the other two are and how we apply that. And then we pick those off and we're going to run with those and truly build out the people process tools at a global level. Because we have to work with our global function, the geography level that's across our Asia Pacific region and then in the countries.

Shaheen Hoda: Sounds super easy.

Craig Mills: Yeah, simple. We'll do it tomorrow and get it done.

Shaheen Hoda: Do you know what the other two might be?

Craig Mills: I have my opinion on it, but I don't want to bias what the rest of the team want to do. So it could be an event process, it could be event webinar process. We do a lot of those. Um, or it could be a sales automation process. I'm leaning towards the sales automation process. Again, I'm looking for the yield, I'm looking for the returns. Productivity is nice, but growth's more important. Yeah, got it. So leaning towards something that's a growth driver. But the problem with the growth driver is that it's harder to do. The productivity stuff's easier. So that's the decision point again is

Shaheen Hoda: you want quick wins or you want solid wins.

Craig Mills: Yeah, well, you want wins that will yield the results in the marketplace. Uh, and I know given from what we discussed as a leadership team on Thursday night, it's growth and it's growth and it's growth. And that's why I'm hesitating to say this because I truly don't want to bias this. I've got my own ideas, but I really want to validate those ideas against the strategy. And that's the key to it. I don't want to pick a use case, for example, that I know is my own personal project, but it's not going to align to Strato. You don't want to align to strategy.

Shaheen Hoda: It excites you. But maybe not necessarily the rest of the or would have the impact for, uh, organizationally.

Craig Mills: Yeah, exactly.

Shaheen Hoda: I have one last question I want to ask you here before we do some rapid fire questions that you have not seen.

Craig Mills: Okay.

Shaheen Hoda: What is your kind of honest advice for B2B marketers who are rolling out AI and really thinking hard about, hey, where, what do I do with this? Right? Um, and I'd like, if it's possible, I'd like to kind of decouple that answer from IBM's capabilities. Right? Because not everybody would have, would be either able to afford or have uh, access to, um, that. What is your Advice? I'm a B2B marketer. I'm like, I'm feeling left behind and you know, I feel like I need to do something. See all these thought leaders on LinkedIn posting about their 6,000 agents, um, you know, running their life and business. Right? What is your, what is your advice?

Craig Mills: Yeah, so we did a piece of research and this has been my North Star for the last probably three months since I think it was released. And it's the Enterprise 2030. It's on the IBM website. Enterprise 2030 interviews of C level execs in enterprise around the world. So business to business work. Okay. And the findings are, uh, pretty simple. 80% roughly know that AI is going to deliver value for them. Only 24% know where that's going to come from. So that's a huge gap. Business leadership worldwide, knowing where they want to go, don't know how to get there. That's one piece. They've called out the five priorities. Product and service, Innovation, productivity, speed of execution, consistent customer experience, AI and technology modernization. M. So that is the C suite of the world's enterprises, laying out the direction of where they want to go. Now for me and marketing, what's marketing's role in that? What's a B2B marketer do? So the B2B marketer, go back to, to your roots. It is customers and partners, number one. Listen to the customers and partners, number two is get closer to product and service development. And we found that internally I found it's incredibly powerful for me. What does that mean? That means you've got to know what you're marketing and a lot of marketers don't actually know that.

Shaheen Hoda: Yes, I've been in those rooms. So what does the product do?

Craig Mills: And the last one, Speed. I call this sort of the triple threat. If you can be, you know, the singing dancing actor. Right, That's a triple threat person, you know, Hugh Jackman style. Right. As a marketer, you gotta be close to your clients and the partners that serve them. If you're in a partner network, B2B is usually like that. You gotta be close to your product and you gotta be close, uh, to the product development team as well to know what's coming and then you gotta be faster than everybody else. The speed of the market is insane. Now that is incredibly tough. But I'm now using that as my navigation point for any time I think of doing anything with AI. Is it going to get me closer to the customer? Is it going to be closer, uh, to the product or service and will it allow me to go faster? And I wake up every day with those three navigation points and everything I do is either in service of the customer partner, uh, getting closer to the product and making me go faster. So Copilot, for example, you know we've got Copilot at work. I use that to just go faster. I'm trying to get my workflow dealing with um, emails and this is internal emails, meetings. And so I'm just doing that. I'm using some of the external tools because it's external research on how to understand the customer better. So I use Claude and uh, whatever else GPT to do that. And then how do I get closer to our product? How do I really understand the product and apply that internally and externally and using tools to do that. So they're my three navigation points. My advice for anybody.

Shaheen Hoda: Such a great advice, such a great point, the three of them. Um, uh, and I think it creates so much clarity because just like you said at the beginning, there is so much noise in the market, um, so many people doing stuff that doesn't necessarily move the needle. It's just cool maybe, but doesn't necessarily move the needle. Before we get into the rapid fire questions, is there anything else that you think is important that maybe. I didn't, I didn't. We didn't touch up.

Craig Mills: Yeah, one piece. And this is something that for those of us who are in charge of teams and people building organizations with maximum mental flexibility, I think is going to be one of those skills and enterprise capabilities that is going to be a premium over the next, whatever the time period is. But certainly next, next one to five years maybe once all this shakeout starts to settle down, maybe it'll be a year, maybe it'll be hard, I don't know. But that maximum mental capability and flexibility to be able to on the one hand understand all the moving parts and then internalize it, stabilize it and then deal with it. I think that at an individual level and at uh, an organizational level is the one thing I would suggest leadership need to look at and how they develop and enable staff to deal with what's coming. That's certainly what I'm doing. How, how would one do that first thing is being clear, clear on the mission. Because if you don't have clarity of a mission, you can't then free your mind to then look at all the variables. And this is what I'm finding really empowering at the moment with IBM is that we have set ourselves a very clear mission. For a long time we didn't really have a clear mission.

Shaheen Hoda: Can you share with that, Ed?

Craig Mills: Yeah, AI, uh, leadership in the world of enterprise. It's really simple. How do we help the world's enterprises become AI driven at scale and with security? That is our mission. So within that, how do we do that? We have to have the best AI products we possibly can and the best AI consultants because we're a consulting business too, to then drive value for our clients. Number one, that's it. It's about the clients and it's about AI for our clients and empowering them to be the best they can be. Now for my particular division, because I'm in the product division, I'm not in the services division, my job is to drive growth for five key products, the last of which, I'll tell you in a moment, it's not a product pitch. We are, uh, just releasing a new product to be a code development assistant.

Shaheen Hoda: It's the software development piece.

Craig Mills: Software development. That's why we will go incredibly hard at the IT space and that other uh, division that I didn't go into before, but I think it's worth going into because it's an illustration of a product designed specifically for enterprise code development. However, it's enterprise code development at scale with security, with guardrails and all the things that people expect from IBM. Now we're launching that. Now it's been in soft launch for a while. It'll come out. That clarity of vision around where that product sits in the marketplace is fantastic. And what that does is then gives us the stability because we know what the navigation point is, we know where we want to go. We then have the flexibility to get there, the permission to get there by not all means necessary because you can't do that. But uh, within our uh, guidelines and our corporate expectations. And that's incredibly powering and empowering for the staff because it allows us to then be flexible. And that uh, maximum uh, mental flexibility of coming up with many, many different ideas on the way to get to that point and to use AI to get to that point of driving AI. That to me is what the necessary ingredient is to empower people to have maximum flexibility and mental flexibility. You've got to provide them with a couple of key navigation points so that they don't get completely lost at sea. That's where we're going to drive to.

Shaheen Hoda: Let's do some rapid fire questions. Okay, so the first one I got is what is one resource that has had a fundamental impact on the way that you work or think?

Craig Mills: Yeah, I've been listening to a lot of jaded Gallicanis. Um, and this week in startups.

Shaheen Hoda: Yeah, right.

Craig Mills: The reason why is that we are the opposite of startup. It's the complete end. And so I wanted to switch my mindset and I found that incredibly useful and very fortuitous because I started listening about a year ago because I thought I need a spark to do something differently. I felt like I was being a bit institutionalized. Been at IBM a while, as it turned out, it was very foresightful, if that's a word, because we just launched this internal startup and I feel like the 12 months of listening and learning and trying to apply and pretending to be a startup within an enterprise, now I'm running. I feel like I've done the training for when the real games hit. So, yeah, that's been one.

Shaheen Hoda: If you can give one piece of advice to B2B marketers, what would it be?

Craig Mills: Get close to the customer and the partner, and don't forget the partner side of things. B2B is different to consumer, and I've never really done consumer, but I've watched it from the outside. But B2B is about partnerships, partnerships with your clients and partnerships with your vendors and the people who work with you and the ecosystem. So don't forget the partners and the ecosystem within your client service mission.

Shaheen Hoda: That's a great, that's a great point. Um, okay, third question, third question. Who are some of the. We talked about the thought leaders of, on, on LinkedIn, talking about agentic movements and all that stuff. But who are some of the people that you kind of look up to or you listen to, you follow in terms of what they're saying? Uh, in general.

Craig Mills: In general. Uh, the one I really like, and this gives me my sort of historical economic, is Ray Dalio. I love the way he frames his forces that shape the world and the historical context. Anybody who wants to see why the world is as it is today, uh,

Shaheen Hoda: there comes out the geopolitics and the history and all that, uh, interest.

Craig Mills: The Dutch, then the British, then the Americans, and now what's next? Uh, that's fascinating.

Shaheen Hoda: What's something that excites you about B2B today?

Craig Mills: I really, really want to ignite. And this is a very IBM specific question. I want to ignite and be part of, uh, a growth division that does more than just 10% growth. We've done 10%. I want to do 50. I want to do 100. We've never done that before.

Shaheen Hoda: I mean, it's a harder to scale of IBM.

Craig Mills: It's massive, right? I want to do 10x. That's what I want to do. That excites me and I feel like we've almost got a product now that can do that. And so I'm really, really pumped about that.

Shaheen Hoda: Go 10x. Craig, I just want to say thank you so much for coming on the podcast. This was an awesome conversation.

Craig Mills: Thanks for having me.

Shaheen Hoda: We hope you enjoyed this episode. If you like APAC's B2B Growth Podcast, please share it with your B2B friends and subscribe for weekly insights on B2B growth across APAC. Sign up for the XG Weekly Newsletter link is in the description and if you'd like to let us know what you think about this episode or in general the podcast or anything else, you can email us@podcastgrowth.com we read all those emails. Apex B2B Growth Podcast is produced and edited by Alexander Hipwell and music is by the mysterious Breakmaster Cylinder. We'll see you next time.