Solving the AI Cold Start Problem - Managing the C-Suite & Culture Gaps | Abhay Parasnis

0:00The CMO Confidential Podcast is a proud member of the I Hear Everything Podcast Network. Looking to launch or scale your podcast? I Hear Everything delivers podcast production, growth, and monetization solutions that transform your words into profit. Ready to give your brand a voice? Then visit IHearEverything.com. Welcome to CMO Confidential, the podcast that takes you inside the drama, decisions, and choices that go with being the head of marketing. Hosted by five-time CMO Mike Linton. In marketing, everything must work seamlessly. If not, efficiency, speed, and ROI all suffer. That's why Quad is obsessed with making sure your marketing machine runs smoothly with less friction and smarter integration.

0:52Better marketing is built on Quad. See how better gets done at www.quad.com slash build better.

1:07Welcome, marketers, advertisers, and those who love them to Chief Marketing Officer Confidential. CMO Confidential is a program that takes you inside the drama, the decisions, and the politics that go with being the head of marketing at any company in what is one of the most scrutinized jobs in the C-suite. I'm Mike Linton, the former Chief Marketing Officer of Best Buy, eBay, Farmers Insurance, and Ancestry.com, here today with my guest, Abe Parasnis. Today's topic, solving the AI cold start problem, managing the C-suite and culture gaps. Now, Abe is the founder and CEO of Typeface, a company that uses AI to automate personalized marketing.

1:50Previously, he was the CTO and CPO of Adobe and also worked at Microsoft and Oracle. He serves on the boards of Dropbox and Schneider Electric, and we invited him on the show since he has a front row seat on companies implementing AI and what's working and what's not. Welcome, Abe. Hey, Mike. Thanks for having me. Excited to be here. This will be fun. All right. Let's start about talking about the cold start issue. Everyone knows AI is a massive game changer, one of the biggest things in our lifetime. But getting things going at the company level is clearly challenging for many businesses, and the payout economics often look to be a bit far down the road.

2:39Tell us what you're seeing on the implementation front in general. Yeah, no, first of all, again, thanks for having me. I'm excited to have the dialogue, Mike. I think a couple of things. First, and you talked to a lot of people in and out of the industry, in the marketing sphere. I would say, first, I'll zoom out a little bit with AI. I mean, obviously there is a massive wave of shift, both technologically, architecturally, but then societal, industry level. So there is no doubt that the change is going to happen. But as you know, often with these changes, especially in Silicon Valley, there's a tendency to somewhat overhype how quickly these technologies will go to the world.

3:17Exactly, exactly. And you know, one of the things I always talk to the team, everyone knows Moore's Law on the Silicon side, like how quickly things are improving. There's another lot less quoted law called Amara's law about technology, which is we tend to overestimate in the short run how some technology change is going to be impactful and profound in the short run, how quickly things will happen. But then we often tend to underestimate how profoundly disruptive it will be in the real long run. And so I think we as humans have a tendency, and especially in the Valley, to kind of think as soon as some breakthrough technology comes along, that it's just going to take over the world.

3:56World GDP will change overnight. Millions of jobs will be lost. Whatever the narrative is. I think the reality is it's more nuanced, as you know. And we'll talk about it. The cold start problem that you alluded to is actually nothing to do with technology. I am a technologist. I'm super excited about what's happening with AI. the technology pace and the breakthrough in the architecture is unbelievable so don't get me wrong I'm super bullish super excited that that is why I'm doing what I'm doing but I think the human element of change and the complexity of the organizations and especially if you go out of the kind of the startup world and go into the real world of big businesses that are sometimes 50 100 year old companies in entrenched industries whether it's retail CPG like you have spent years in, it's not just about technology.

4:46They have to rethink their business process. They have to rethink their organizations. They have to skill their orgs. Well, and this is true about everything. When you look at digital, and like I was in retail, when digital really kind of was taking off, yes, it's super impactful and changes everything. But there's lots of executional, compensation, structural, other issues that are all in the way. And Silicon Valley does minimize that and will say, if you're not vibe coding, you're nowhere. Yeah. Um, but, but that's not really what, what is the real issue with the cold start problem. Um, but, but there's probably a second issue on this.

5:23If it, correct me if I'm wrong, the C-suite and the board is probably saying, oh my God, what are you doing with this? You got to be on it because you're really, you're not vibe coding. Oh my God, I vibe coded last night. Um, you know, we got to do this for the company.

5:42and the practitioners are probably having a tough time keeping up at least that's what we hear from everybody yeah is that gap for real how big is it it is what are you seeing about it yeah it is a real gap and I'll kind of try to give you as much as I can a balanced view on both sides I don't think this is a issue on one or the other side as in the c-suite or the practitioner I actually you think there are changes required on both sides? So first of all, you're absolutely correct in that every C-suite conversation, I'm at least getting involved in also, as you mentioned, I'm a board member. So I see that both sides, the level of conviction and excitement, sometimes justified, sometimes maybe not, is extremely high about how AI is going to profoundly change any industry, whether we are an energy industry or retail, CPG, electronics, travel, hospitality.

6:31I think the C-suite and the boards are extremely paranoid at some level that they don't want AI to disrupt their business at some level. Exactly. I mean, you're super paranoid. And then between the media and everybody in Silicon Valley and all the consulting companies saying, oh my God, you got to be here or you're going to get annihilated next year. That's the reality. And every day you see hundreds of billions of dollars of investment going into data centers and AI stack. And then if you're a CEO, you're like, oh, others must be on this journey way ahead. How am I not there? So I think the C-suite pressure, excitement, interest, energy around this is off the charts.

7:08I mean, as you have done, I've been fortunate enough to be part of cloud transformation, mobile in the last decade or so. This is different. The intensity and the conviction level that I see at the highest levels of the org to adopt this is very high. Now, the challenge is they don't know what to do about it. I mean, just spinning off a wipe coding or experimentation or some innovation project, you will get a hundred projects like that go off and within six months they'll die and nothing will come out of it. Then the C-suite is like, what happened? Why is my business not fundamentally different?

7:40Why is my customer experience still stuck in dark ages? And so before we go to practitioners, I think the C-suite, and we'll get into this, they have a responsibility not just to be excited, but to actually guide the organization around what is the business process change? and what is the playbook needed? Because if you just say, I want AI and ask practitioners, they are going to be there in the headlights because they don't know what to do and they are not empowered to drive the core business process change, even if they adopted a technology. And where, by the way, the first generation co-pilots didn't help is the narrative got created.

8:15Oh, just adopt this co-pilot. Every employee will be X percent more productive. All right. And you will get benefits of it. Just practice with it, play with it, invest in it. It will all work out. You'll figure it out. And I think that is, well, what we're seeing at least is a lot of people doing it on the efficiency front, but nobody actually, it's really hard actually to do it on the structural and the top line growth, like efficiency. I think there will be efficiency, productivity gains out of AI, but the real play is how do you use it as a competitive advantage at the top line growth level and what we are seeing is two very different kinds of companies that we work with companies that only are chasing some cost saving let me drive some efficiency cut some people here yeah i mean that may happen and some jobs will shift but the ones that are saying let me use this moment and this shift to rethink how i reach customers or how i engage with my brand and fundamentally drive few points of better mid-funnel performance in my marketing or how do i drive revenue growth.

9:19So that's on the C-suite. You ask the other side, the practitioners. Well, there's another thing, which is it's so much easier said than done to say, oh, well, we'll just re-engineer around this, you know, brand new thing where there's nobody's actually been working with it that hard for a long time. And you just go do that. And we're on the board. You tell somebody and say six months report back whether you change. Tell us how you changed the company in six months. Here is a quick word from our sponsor, Quad.

10:16Better marketing is built on quad. Now back to our discussion with Typeface CEO, Abhay Parasnis. You know, I want to go into kind of how this thing unfolds as people are taking on the cold start problem and how to recognize the blockers in the cold start problem. But, you know, you guys shared some data that said content demand is exploding. and we hear that from a lot of our listeners. Agency spend is usually under a lot of pressure and what you hear from people is, I want to do this, but I don't know if I can keep up with the pace. I look at this and I say, gosh, you can see a year down the road where plans and forecasts are going to be out there assuming all of this works and they get missed and that a bunch of people are really mad.

11:13Is that what's going on? Yeah, I mean, I think that is certainly a component of it. Your comment around the pressure on doing more content. And by the way, as the world of AI and agents and these answer engines actually take over more prevalently in our lives as consumers, the demand for even more nuanced storytelling and content, it's no longer enough. I mean, I'll just give you one data point. I was with a lot of customers last week in New York. and some of the biggest brands, one in financial services and insurance industry, one in media, they both said, the CMOs, they just ran an audit and already 50 % of the traffic to their website, for one it was 50, for it was 40, is no longer coming from human users coming and reading the website.

11:56Right, it's coming from bots. It's agents. And so you think about, what does that world even mean? How do you reach your actual end user with a compelling content if those users are not even coming to your website. They are consuming it to some other mechanism. And so I think the pressure on content that you said is very real. The gap between expectations, forecast versus the reality. I will say, Mike, there's maybe a little bit of a dual story here. It is widening in some cases. And if you are not careful about not just setting the metrics out there, sometimes I talk to our team that there is a phenomenal in a lot of functions, but in marketing of these kind of metrics that look kind of green from the outside, but nobody's really going deep and looking inside that they are really red.

12:42Everyone stares at dashboards. I talk about them as kind of watermelon metrics. They look green, but they really are not. You have to go deep inside your business. And so what's happening is people will set goals, then they'll blame it on technology or AI. But what they won't realize is, hey, did you actually rethink your process? Did you make fundamental changes or did you just slap this technology on top? And by the way, the role is also on vendors. I would say typeface and others. We have to have that nuanced dialogue with our customers. Look, one of the things I would love to engage with you, you have done this.

13:16The difference in AI, Mike, versus other waves, even technology vendors have to play a consultative role around business process change. This is not just... Oh, I agree. And I also think there's no vendor out there that can tell you the truth. They also, they can't say, we don't kind of really know exactly what to do, but we have an idea. They have to act like they are doing it or they're out of the process. Yeah, I mean, there is certainly pressure. But the one thing I was going to say where there is a little bit of a shining kind of a bright light that I'm starting to see. There are cases where customers and C-suite that say, you know what?

13:54Yes, AI will be eventually profound, but let's pick one narrow use case or one part of my business. and let's put a very surgical laser focus on not just bringing technology, but rewiring the organization. And maybe we can't boil the ocean entire org, but we'll pick one or two processes. We will fundamentally rethink, make people changes, make technology changes, and let's deliver such a compelling ROI that it becomes a lightning rod for the organization. And it becomes a best practice for the org. Exactly. This is the measure. This is how you do it. You can also people handle the issues. So you can, you don't have to, if you do it across the whole org, there's resistance everywhere.

14:37It is. It is. And I'll give you without kind of naming the brand. We are very careful about their sensitivity, privacy. In media, telecom industry, a very global brand, all of us would recognize. They went through this journey with us. And one of the things they did really well, I think it's the best practice in my opinion, is they said, you know what? We have a diversified business, many business units and stuff, but we are going to pick one use case, email, and how we engage our customers, existing customers, and new customers we want to acquire. And we will measure ruthlessly the before and after of email personalization.

15:15When we send emails that are not really personalized and just blast that we all experience, versus use AI to really do extremely fine-grained audience product-based personalization. and they saw, Mike, within first 120 days with us, around 93 % improvement. And this is to me like the company-wide public A-B test. Exactly. Everyone can see all the data, but it also allows you to manage all the resistance and everyone that wants to resist is usually not in the test because I want to move to the resistance process and why. and why people resist, because a lot of it is actually pretty logical and everybody resists this kind of thing.

16:02Not everybody, but a bunch of people resist it. You guys broke it into three camps and the fear of job loss, fear of the new thing, and agencies fighting to hold on to their business. I want to talk about each one of these and how you recognize them, but also how you recognize if you are actually the face of the resistance. Because no one is actually out there saying, gosh, I don't want to do this because I just hate it. They always have a logical reason. And in things like this, where the Martech stack has burned so many people for so much money, a lot of people are going to say, here we go again, even though I think it's totally different.

16:52I want to talk about your big three things, fear of loss, fear of the new thing. And let's start with fear of loss. Java. Yeah, I think, look, first of all, as we go into that, you correctly said, I think in marketing technology in particular, it hasn't helped over the last decade or so. Vendors have promised digital transformation nirvana, hyper-personalization nirvana and they they have it delivered so in a way there's already a little bit of a baked in kind of a resistance which is nothing to do with AI which you've got this gold-plated martex stack that is producing honda civics or whatever i mean basically you know i talked about this before things like cdps the oh as soon as you get a customer 360 profile you will get a full 360 view of your customers and in real time you can engage with them in the right moment right time, right product, it's still a holy grail, hasn't really happened.

17:44And so I think one, I think we should recognize that's the baseline from with a lot of these companies and practitioners start from, which is, hey, this hasn't delivered before. Why is this different? Number one. But I would say the job loss fear, the first one you wanted to talk about, look, what is happening is we as consumers and even these practitioners that may have a fear in their personal lives, they are using ChatGPD, they are using Gemini or whatever. So on one hand, they are seeing the power of these models, whether they are using it, their kids are already adopting them in their college or schoolwork.

18:15So they see the wave coming. But then the fear is, hey, if these AI models are going to do more and more and more of my job, does that mean that by me adopting this, like one of the examples, one of our customers kind of in the retail industry, it was a very genuine conversation. The CMO was 100 % excited, brought us in. The head of creative, really amazing woman, a lot of experience, creative space for the last couple of decades. She saw the demo of our product and she said, look, Abit, can I give a real honest feedback? The product is amazing, very impressive, crazy how good it's getting, the AI.

18:52But if I bring this into my team, what's to stop my team from getting cut in half in the next six months? Nothing really, probably. And by the way, interestingly, She was at least honest about it with us that like, look, I'm not sure I want to aggressively bring this in, but I will go along because the CMO was asking. And the reality, Mike, six months later, her team did get, unfortunately, some budget pressures, did actually get asked to reduce or rebalance. Then she called us and said, you know what? But if this is the inevitability of the business where I'm going to get asked to do more content at faster pace with lower budgets, these AI tools may be the only way I can do it.

19:31Well, and if you don't do it, your cost structure will put you at risk of losing those jobs anyways. She's actually become a champion at that level. So what do you do about it? You have to explain it? Because you're going to have, and it's justified fear. Any justified fear. By the way, I don't think we should just say, oh, practitioners don't get it. One is justified fear. I think second, I think one thing we try to explain, and not explain with words, but by going through a use case, like you were saying earlier, is, you know what's going to happen? this may seem like a little bit of an interesting analogy.

20:02When Uber came along, everyone was like, oh, this is democratization. Everything is done at the base level. And at some level, yes, there was a little bit of a race to the bottom or compression on kind of... The way I think about AI, it's going to seem a little bit like a weird analogy to jump from Uber. We love those on this show. I'm a big Formula One guy, me and my son. I don't know if you're a Formula One fan or not. You know, in Formula One, the cars are so unbelievably powerful and engineered to the ultimate performance, where the craft becomes who can master how to actually take these massively powerful engines and get the peak performance out.

20:42And I look at AI as that, where AI is going to get more and more and more powerful. And the craft of actually mastering these tools and extracting the performance, extracting the creativity out of those, everyone will have these tools. And yes, there's fear. So what we try to tell these practitioners that, look, this is going to free you up from the road parts of the job. And yes, you will have to learn new things. But the craft and the race to the top is ultimately going to become the game. If you really become now world-class using these tools at craft of storytelling, brand engagement, how do you stand out?

21:19The irony in marketing, in my opinion, with AI is the more AI democratizes content creation and scale and personalization, the more premium will put on those unique individuals who can actually do the analog. Because you want better, faster content at lower cost. And in the end, you will win. And so that's an inevitable one, but good leaders have to explain that in a way that makes sense. I want to talk about the other thing, just the general fear of the new thing. And we, I alluded to it in the, in the MarTech stack, which is, you know, yeah, everyone had to buy the new thing. Everyone was like, you got to buy it, you got to buy it, you got to buy it.

22:00You were one of those companies, you got to buy it. Used to be in them. So that fear of the new thing, what does it really look like? Like, and I think you maintain it's different than the fear of the job loss. It's just, I don't want to do it. Tell us what it is and why. Yeah, I think there are two elements. And in MarTech, there are two elements of this fear of new thing. First, I think that, as you know, in the way marketing functions have evolved, not even MarTech stacks, but the function, there is an aspect of accountability models that are a little bit different today. As you know, there's a lot of outsourcing that has happened over the last decade.

22:35Agencies take on certain portions of your marketing job. And if you are a mid-level marketing practitioner, a lot of times your role in some ways, I'm somewhat simplifying, oversimplifying, is outsourcing to some agency, holding them accountable to do your marketing campaigns, measuring if you can to begin with, how good or bad their performance is. Exactly. What AI is now going to do is going to shift that accountability. So it's not just fear of learning new thing. It's going to bring the accountability back to those practitioners and middle level like, hey, you now have the ownership of those outcomes and you will be, you're not just able to point to an agency and say, oh, they did it and it didn't work.

23:17Can you give me a good example of where that's manifesting? Oh, in terms of how they show? Like one example, I mean, there are a few things. One, there are lots of people who will run their social media ad campaigns and stuff, but the efficacy of what creative and what content actually is performing. Today, that is such a black art. Nobody knows. Like you give it to agencies, agency will do a bunch of billable hours around resizing banners, doing these, publish the campaigns, and then you get some disjointed reports, whether it's Google Analytics or from Meta or whatever. and if you are a CMO that says, can somebody draw me a correlation of which campaign?

23:54Yeah, between this and saying sales. Yeah, nobody can do it. And so there what happens now, eventually the answer is going to shift from, oh, you can't just say, I will ask my agency. You will have to explain with these new tools. So I think there is a little bit of a learning curve aspect, but there is also an accountability shift that I do think is a genuine fear. So let's move on to the third one, which is the agencies don't want to give up this business and the consultants don't want to give up this business. So they're going to tell you they're on the cutting edge of AI.

24:30How do you even know? And this is a real problem because if you don't really know AI and they're telling you they know AI, there's a lot of problem in this relationship because no one's going to say, we think we're on the front end, but we don't really know it that well. Yeah, there's a lot of noise, as you said, because everyone is saying that technology vendors are saying they know it and they have solved it. By the way, I will first of all say agencies do play a very critical role and I think will continue to play a role if and how they evolve with this technology. Like there is a change management piece for CMOs and practitioner.

25:09There's a massive change management piece on the agency side. So I don't think this is like as simple as saying, oh, agencies are dead and now everything is going to be done by AI. I think that's naive as well. But there are two things fundamentally I do think agencies have to recognize. The notion of a model where it's just billable hours and you keep doing repeated kind of brute force jobs. I think they are going to have to be able to demonstrate that they can do more higher value creative jobs because the base tools will now do at least half if not 70-80 % of the job that they used to be able to charge billable hours.

25:43So one, I think, is the recognition that you can keep telling your clients the story that nothing is changing, and maybe that will work this year, maybe it will work next year, but the dam is going to break. The dam is breaking. Well, the cost structure, in the end, the cost structure is not going to be sustainable. And everyone knows it. You alluded, we just published a report last week called Typeface Signal Report. And even these early days of AI, the report we interviewed hundreds of CMOs and leaders, and they are saying 60 % of them already are saying they have decided strategically they are on a declining agency spend path.

26:19And it's just a pace and how they operationally execute is the question, not if they're going to do it, is how they're going to do it. And so I think - But only 14 % of them said they thought they could keep up. That is the other piece, which is the pace of content creation. And so they are fighting this dilemma. I do think, by the way, agencies, some incumbents, but some new ones that are starting to recognize this will emerge. Well, if you're AI native, you don't have to deal with all the I think that's the thing. You can adopt the AI native approach. You can adopt these technologies, but then you focus on, you know, one of the things kind of I was going to say is this, I'm sure you have heard of Malcolm Gadwell's kind of 10 ,000 hour kind of thing about what it takes.