Welcome to the Marketing AI SparkCast, where AI meets modern marketing. Hosted by Aby Varma, a global leader in marketing, AI, and innovation, each episode brings you real conversations with leaders putting AI to work in strategy and execution. This episode is brought to you by Spark Novus and the Marketing AI Pulse community. Here's your host, Aby Varma. Welcome, everybody, to another episode of the Marketing AI Pulse brief. This is the March 2026 episode. For those of you who are not familiar with this format, this is an extension of the Marketing AI Pulse community, which is one of the fastest growing communities focused at the intersection of marketing and AI like the name suggests. To learn more and to join for free, you can go to marketingaipulse.com. And as always, I am super excited to welcome my partner in crime, Matt Cyr. Matt, welcome to the show. Happy. How how are we doing today? We good? We're doing good, man. I can't believe this is our third one. And for our listeners, Matt and me do this every month where we'll curate the most relevant stories for marketing folks so you can you can get a better understanding of what is happening in the landscape. Matt, how are you doing, man? I know last time we spoke, you were not doing so good, but everything good now? Yep. I'm good. And I'm not nearly as doped up on NyQuil as I was last time, so I should be I should be raring to go. You ready to kick us off? What, I know things are usually pretty quiet in the world of AI, but anything happened last week we should talk about? Matt, yes. The AI world was quiet not as always. So this week, we had NVIDIA's GTC conference, and Jensen Huang introduced something super cool. He always does. He always has exciting and shiny toys to announce, and he's really driving the infrastructure layer in the world of AI. For those of you not familiar with it, Jensen Huang is the founder and CEO of NVIDIA. That's the company that started off doing gaming chips back in the day and now has got crazy valuations really behind the infrastructure for a lot of these AI crunching, the chips that do a lot of the the AI crunching. So really infrastructure layer. And he has been innovating or NVIDIA has been, like, really innovating at breakneck speed. And what he announced this time was two layers in the world of AI. So one was a build layer, like how do you design and create AI agents, and then a run layer that how do those agents operate inside your business. And why is that important? How do you design and build AI agents? There is no common way of doing that today. Everybody does it differently. Microsoft does it differently. Google does it differently. You have some of these application providers like the Salesforce's and Adobe's and SAP's. They all have their own methodology of doing it. So what NVIDIA and Jensen are doing is really providing the infrastructure for application folks to start building it and running it and doing it in a manner that is consistent. To understand the BuildAir, they call it the agent toolkit. And what it is essentially an open platform that helps enterprises design and develop agents. Right? So this is not a new feature. It really reflects a shift towards a standardized infrastructure and how agents are created, like I said. And the significance to me lies in the ecosystem. So they... He announced loads of companies like the the Adobe, Salesforce, SAP, Cisco, ServiceNow. They're all working with NVIDIA. There were more, I think, but they're all working with NVIDIA to bring these capabilities into the platforms that people are using every day already and marketing folks as well. We have Adobe and Salesforce in there as well in SAP. So the signal... This signal is basically where the stack is heading, where we're really getting a vocabulary, almost like a shared way to build agents across the tools. And what does that really mean? It means that we are moving from chatting to doing. We already knew that, generative AI to agentic AI, but now there there is a way... A standardized way to approach that. In this build layer, aka the agent kit, there are three things, and some of this sounds super technical, but I'm gonna try my best to simplify this for our listeners. They have within the agent kit, there are three elements. So there is reasoning, which they call it Nemotron. Sounds like a video game name, but whatever. The models that allow agents to link through kinda multistep tasks, like campaign planning and optimization and that sort of thing. So that's one. Then there's an element of security, which is always the concern for enterprises, and they call that OpenShell. So this puts into place guardrails that ensure agents follow rules around privacy and compliance and that sort of thing. And then, AI-Q, which is efficiency, an approach that really makes agents faster and more cost effective to run at scale. So these are the three elements of it, and that was the build layer. And then you have the run layer. It's called NemoClaw. So, you know, we've all heard of OpenClaw. Finding Nemo, Aby. There's, you know Yeah. Right. All of these together here. Seriously, man. These names are, like, too cute for what they do. But... Yeah. So NemoClaw is... NVIDIA introduced that, and that represents the run layer. And that builds on OpenClaw, the concept evolving it into really an enterprise ready platform designed to run and provide governance to AI agents in real world environments. Right? So, essentially, if the agent toolkit is how you design the system, NemoClaw is what makes it work in the real world. So it runs and coordinates agents across systems. It manages workflows and execution. It adds that enterprise grade security control and scalability, all of which was sort of missing in OpenClaw where people were freaking out about getting hacked and all that kind of stuff. So a lot of technical stuff, but what does this really mean? It's pretty major where you can actually make AI do work. Right? And, essentially, what's gonna happen is once this will manifest itself in the form of business application, then you'll suddenly have the, you know, the Adobe's and the Salesforce's and the other application providers starting to embrace this framework. You're gonna move to a tool that assists marketers to one that participates in the actual execution of work. Right? So... Which is pretty powerful. Right? So NVIDIA is positioning itself as kind of a foundational layer of AI on how agents are built, governed, and scaled across the enterprise stack. And for marketing leaders, that shift is clear to me, where the question is not just what AI can create, it's what AI can actually run. So before, there was a lot of discussion about, hey. Can... You know, we we want to make sure that the generative AI is creating stuff, but there still... There was a lot of work that needed to be done. Sure. You can use AI to go generate a blog, but then somebody's gotta... Obviously, you want a human in the loop, but somebody's gotta take it and post it on your website and do other things to it and push it out on social media and maybe in your newsletter, whatever. You know, whatever are your derivative actions after that initial thing, which is all... That you could mitigate some of those manual things using automation tools like a, you know, like a Zapier or an n8n or a Make or what have you. But now with some of these capabilities, I know they just got announced, but with some of these capabilities, it's seeming more and more real that it is moving from creating just to doing. So I don't know. Has that blown your mind, did you get lost in the nomenclature of Nemotron and NemoClaw? I think Nemotron was a transformer. Right? I'm pretty sure. Sounds like that, man. Sounds like that. This tracks with not only everything I think that that we're seeing, but actually working on building a platform right now specific for marketing agencies that's really built on agentic structure, right? And really training these agents to go out and do the work wherever possible on behalf of the humans when there's a task that is repetitive, process oriented, all of that kind of good stuff. So I think to me, the biggest sort of moment from this is that it's coming from NVIDIA and Jensen Huang, And to your point, it's not just a shift. This is a massive decision by a massive company that has enormous clout in the world of AI saying Yeah. Agents agents are a thing. You may have thought they were just a passing fad, but they are a thing, and let's show you what we mean by that. So I think it's great. I think it just doubles down on the folks who have been saying for a bit that agents are gonna be an important part of the future. I think it's really cool. I think it's really cool. Yeah. I mean, I like the fact that this is giving the industry the, you know, the the tools and the know how of accelerating, building, and running agents in a safe way. Yeah. So to me, that's a huge leap instead of every software company trying to figure it out themselves. Well, here you go. Now there's kind of infrastructure in a way to build it so you don't have to figure it out because NVIDIA already has. Yeah. Yeah. And it's... It moves us even farther beyond a lot of companies and tools were built as LLM wrappers. I'm gonna take ChatGPT, and I'm gonna build a business around it. I think that trend really is, in a lot of ways, coming to an end. And I think what we're seeing here is probably the next wave, which is really around agentic behavior and massive infrastructure to your point. Good stuff. Yep. Absolutely. So we move from this kinda super innovative future forward to what is happening now. So our next story, Matt, is the fact that there's information and and data, which talks about how true AI adoption is still lagging in marketing. Although, I see the state of AI in marketing by Salesforce and all these different reports claim that, you know, you know, 80, 90%, whatever the exact number is, of marketers are actually using AI. But I think based on some new data, it sounds like that is not the case. So can you unpack that for us? Yeah, so it's interesting, yes. So we went from very big picture NVIDIA Jensen Huang AgenTeq systems to the other end of that, which really lagging adoption, if you will. Supermetrics, which is a marketing intelligence company, released a survey report where they found that only 6% of marketers said that AI was fully implemented in their marketing operations. So not the sort of 80 to 90% of we're using it, we're trying things, we're experimenting. 6% saying fully implemented, it's running our marketing operation. And again, this is a study from a company that has a dog in the fight, so we can probably look at some of these data points with a little bit of skepticism, but I would say the reason that I wanted to talk about it this week is primarily because I think that the survey results track with what a lot of what I'm seeing out in the marketplace, and I think you are as well, when we talk to people who are adopting AI. So again, it's at 6% of marketers, is fully implemented in their marketing op, 80% of marketers are feeling pressure to adopt AI, and 37% said they lacked a clear AI strategy from leadership. Right? So on the one hand, have leaders pushing to get AI implemented. Their people are saying, yeah, but we don't really feel like there's a strategy here. We're just running in circles. And then 6% saying, yeah, we've actually fully implemented marketing across our ops. I actually would push back on that 6% unless it's a group of companies that you and I just never see, because I don't know anyone who is fully done with their AI implementation across any industry or any type of company. So I would look closely at that number as well. But I think what this tells me is, again, what we see out of the market is that people are really finding AI adoption right now to be what I think of as death by a thousand cuts, right? This group is going to try this set of tools, the other group is going to try a different set of tools, the two groups are not going to talk to each other, there's not going be a formalized plan around what tools we're using for what purposes, and how are we measuring the impact of that work. It's just in a lot of ways very scattershot. Now, I don't want to undercut that, because there is a value in experimentation as you and I have both recommended to our clients, but also obviously you and I do every day in our work lives. So there's value in experimentation, but I think we're moving very quickly to the point where experimentation has got to become focused implementation. And we have to move from this death by a thousand cuts mentality to a place of we have a plan, we have a strategy, we have a plan, and we're going to run that plan and measure that plan, or measure the KPIs of that plan and make adjustments based on that. So to me, this is really indicative of this, I think, important moment that we're at in AI, which is people are doing a lot of things, they're trying things, they're getting increasingly comfortable with it, but it hasn't yet really become the function that has been decided upon and is driving business. Do you feel like you see the same thing in your world? Yeah, for sure. I certainly feel that there's a lot of AI usage, but yes, in terms of... Firstly, there is no even definition of fully implemented. I don't know. I've not read the Supermetrics report, so I don't know if they No. They don't offer they don't offer a rec... A definition. So Yeah. Leave it up to us. Yeah. But I think even, you know, for folks that we speak to, you know, as part of the Marketing AI Pulse community, there's a lot of community exchange and that sort of stuff. Even that is really interesting because people will say... In fact, I was talking to a Fortune 100 marketing executive, and and he was like, yep. We use AI in our workflows today and everything. And, you know, I was like, how do you guys use it? What exactly do you do? And his interpretation of that was notetakers. And I'm like, okay. So Yeah. I think that, you know, using notetakers... And I'm sure we didn't get into the weeds, but I'm sure I... I'm I'm thinking that the notetakers are used to get insights from subject matter experts, which then get converted into content somehow. I'm thinking that's what he meant. But, yeah, just use of notetakers does not constitute where, you know, using AI in marketing per se. But, yes, I think there is no standard definition of what this is and what is, you know, what is the right level of AI being embedded in your marketing function in order to claim that you are fully, quote, unquote, implemented. But I think, certainly, I think there's lots of experimentation, a lot of playing around, you know, a lot of ideation and content usage and writing emails and those sort of things. But, like, embedded, you know, using marketing as, using AI as a marketing operating system throughout all marketing functions and everything, definitely not even close. Yeah. Agreed. Not even close. And it's interesting, I was talking to a client a couple weeks ago, and this is a company that has engaged us to do some AI consulting for them, really help them do some of what we're talking about here, which is put this into your workflows, into your operations, make it more of a bedrock foundational piece of the business. They're investing time, energy and money in this, they've got leadership's attention, they're doing all the things that you and I would recommend people do, and yet they tell us all the time, gosh, we feel like we're behind, and everybody that I talk to says the same thing. And so what I always like to say to folks is, look, the fact that you're engaging with us in this work, the fact that you're thinking about it in this way, the fact that you're experimenting in different ways means that in a lot of ways you're way ahead of the vast majority of the folks out there, and so give yourself some credit for doing the work that you're doing. And I would actually say to build on that in thinking about this particular survey is find ways to really start to operationalize and systematize this, right? It's, yes, we've talked about two to three workflows, find workflows, find challenges in your business that can probably be well handled by AI, and really make AI a part of that, force AI into those processes so that you can start to see where there's gonna be some benefit and some value, and of course, measure the benefit that you can get from that and learn and evolve just like any other project. I think one of the things that we see a lot is lack of governance in these organizations, even the ones that are committed to it, and I think there's a little bit of the we wanna wait for the perfect moment to do this governance stuff, and so I would just tell people don't let the perfect be the enemy of the good. Identify a person or a team to really lead these efforts, have responsibility for these efforts, codify some of your governance, right, put some policies and procedures in place, those are going to change, everybody knows that, AI is changing too fast for this stuff to stay around for a year or more, but start to codify some stuff and get things, give your people some guardrails and some guidance so that they can move forward with some clarity. And finally, would say a lot of what I I see marketing agencies, marketing organizations in particular doing, is focusing on the cool shiny stuff. Oh, have you seen the new video generation tool or the new SEO tool that does this cool thing? That stuff's fine, it's part of the experimentation layer, but I would recommend that folks take that strategic step back and say, let's look at this from a business perspective and figure out where the greatest challenges and then thus value are coming from. So take that 6%, I would say not as, oh my gosh, there's already six percent of companies that have already completely figured this out because I don't think there are, but say there's a huge opportunity for us here within those numbers to really push forward in a way that's going to give us both market penetration, but also competitive competitive value. Yeah. Couldn't agree more. I think the one thing we just said, which really resonated with me, and that's our advice to clients is that you've got to attach your AI initiatives in marketing to some sort of a business value. Otherwise, it's just an experiment that is going to eventually wither away. Like, I see a lot of companies, like you said, use tools, and they'll play around with it and that sort of thing and be like, yep. This is great. It works. But then that's it. That's the end of it. Right? Because there is no use case. It's not tied tied to business value. You're not getting into it with an understanding that, hey, this is a larger initiative other than just quote, unquote, a tool. And without really doing that, you're not gonna go anywhere. So I think the attaching it to that business value is a very relevant point you made. Yeah. Absolutely. Alright, Aby. I think you wanna talk about agents again. This is the a... The Aby agent show today. Gonna... I'm gonna ping pong this back across the table, and you're gonna tell us a little bit about FedEx Yeah. As agents, and now you're gonna about FedEx as agents. Yes. Absolutely. And and this one caught my eye. This was like a Wall Street Journal article from last week, I believe, which was really talking about FedEx is building an AI agentic system, you know, designed to really take their business objectives and translate that into coordinated execution across functions, not just marketing, like across functions. However, the piece specifically called out marketing operations and campaign management as one of the areas where the system will be applied, which I thought was very interesting. And there, they... The article spoke about, you know, the... Their structure of the agents was sort of three elements. Right? The manager agents the manager agent that structures the process by breaking down the goal, like launching a campaign and breaking it down into sort of baby steps along the way. You have the worker agents that carry on those tasks across the system, such as data or messaging or channel execution or what have you. And then you have an audit agent that review... Reviews the outputs for accuracy and compliance and quality. So to me, I just thought that this was the fact that this... They call out marketing and the fact that this is yet another proof point that the world of agents is live and well and it is going to be more and more prevalent. I think it's starting small, but we can already see that it is infiltrating all aspects of marketing and is going to continue to do so in the future. One wonders that in, like, ten years, is it gonna be lean and mean marketing departments with a whole bunch of agents doing the execution and the marketing folks doing the the thinking and the strategizing, but then the actual execution is going to be done by agents. And, yes, there may be a human on the loop in the other side, you know, of that pipeline to who is gonna review things. Although this one has an audit agent, but just wonders what the future marketing department and operating model would kinda look like. And which is why this sort of caught my eye because I felt that, you know, for marketing leaders, there's three elements that stood out. The operating model, like I said, so marketing execution paradigm shifting from, you know, managing workflows to really designing how workflows through the systems, right, where you wanna make sure that it's not it's not humans connecting. It's not a Zapier make any type workflow thing, but it's more you are imagining and directing how you want your agent to execute without really manually connecting things at every single endpoint. The execution impact, where campaign execution becomes more of a agentic coordination effort end to end, reducing manual handoffs and enabling faster movement from objective to launch to optimization. And then finally, governance. You know, they... Like I said, they have an audit agent, so brand compliance and governance and quality checks are embedded into the process, which is why, again, the governance thing that you were talking about earlier, the policy, the governance policy would be super useful because what if every marketing department had governance, and then you make sure that from from an agentic standpoint, that information is not codified and your agent can then understand that and has some sort of an evaluation rubric against what it actually does. So I think Yeah. It's early days, but certainly, feel that CMOs and marketing leaders should start thinking beyond the tools and focus on how campaign workflows could be redesigned to incorporate, like, system led coordination and built in governance. Again, it's not commonly accessible. Again, FedEx is a huge company, and a lot of these tools are now enterprise companies, but it's just a matter of time where I think these are gonna trickle down to where this is going to be the way. This is the default way that marketing is going to be done moving forward. That is. You know, and to build on your governance point a moment ago, I think the kind of conversation that I think marketers are not yet having, that they're going to need to start having is, what are we allowing AI agents to do on their own? Right? What the keys to which cars are we gonna give them? Because these agents, to your point earlier, are capable to some greater or lesser degree today, but certainly will be more so in the future, capable of actually doing things beginning to end that may not need a human to do anything. And I think that raises a really important question around, okay, what do you want, what control do you want to give over to the AI, and where do you want to make sure that human being comes in and has a decision point that says yes or no to a recommendation that an AI agent is making? There are gonna be certain things that some companies are gonna be very comfortable just letting them run and do certain things without a human touching them, but there's other things that are obviously really business critical that if there's not a set rule for the how the agent operates and when it asks for help, and when it asks for sign off, you could see companies getting themselves into some significant tough corners. So earlier those conversations happen, in my mind, the better. Yeah. And I think it's so it's so new. Obviously, the bigger the enterprise companies have access to some of these things, like I said. But I think for a normal, you know, even mid market marketing department where some of these things are not entirely accessible, I think there is still room for marketing leaders to start thinking about this stuff. And if you're in middle management and marketing, I mean, this is going to be your reality by the time you were going to be in the world of, you know, senior leadership and marketing. And if you're, you know, heading towards the the CMO as as your destination, then I think certainly start educating yourself as to how the impact on on marketing orgs and the impact on workflow and the impact on governance and all of these things. I think it it can't be ignored, essentially. Like, it is gonna come... It is coming to you fast and furious. Yeah. Yep. Sure is. Alright, Matt. We can't have an AI marketing conversation without really talking about Google. Yeah. So tell me, I know there's some new data by BrightEdge about Google and AI. Unpack that for us. Yeah. You got it. Yeah. So yet... So another done by a company. So, again, you can probably look at some of these results with a little bit of cynicism, but I think this one to me, what I really like what the data point says because it has very clear implications for marketers and content marketers in particular. As you said, BrightEdge recently released a report where they did a study of traffic and the kinds of information that ends up in Google AI overviews and ChatGPT AI overviews, etcetera. And the most compelling data point that they found was that Google's AI overviews are 44% more likely to display negative information about a brand than ChatGPT. So just to put that in a little bit of context for you, that means that for every 1,000,000 queries, about 23,000 might yield a negative response by a Google AI overview about your brand, right? So you think about all the work that people are doing to build their reputation online, and make sure that they're showing up well in the search results in and out of LLMs, and oh, by the way, the most powerful search tool on the planet is putting you at risk of having negative responses about your brand be the first response that people see. So you might say, hey, why is why would that be happening? Is Google trying to tip the scale in one direction or another? And the short answer is, I'm sure there's a lot of factors, but one of the things that BrightEdge pulls out is that because of the way that Google sources information for its overviews, it will often use very historical information from websites and from other marketing materials that inform a more comprehensive answer. They're looking to tell a more rounded story, and so BrightEdge is suggesting, or the CEO of BrightEdge is suggesting here is that because of the way Google pulls information, in order to try to give people a more rounded response, it will sometimes pull data from old sites and old information or even negative reviews that people have left about that brand online. And so it's starting to tell a story of search is not just something that you as a brand can control a 100%, it is now very multifactor in what gets put in front of people. And I think it's a reflection of how people are using AI tools to find information. I think you and I live this, and I think a lot of people are living this, which is increasingly using these tools as deep research tools, right? Really helping tell or find information that goes a layer deeper and is not just the marketing ad that somebody wants to have make sure that I see, it's going deeper than that. And what they are suggesting here is that marketers need to be really cognizant of the kind of information that's already out there and the kind of information that they are creating net new, Right? So we all know search traffic patterns are changing. I've heard from lots of people that the traffic on their brand websites is going down. That's to be expected given the use of LLMs as essentially search engines on steroids. So you have to think differently about SEO. Authority is gonna matter more than ever because people are looking for those deeper answers. And interestingly, the data points that we're starting to really look at are becoming... Are changing, right? Where it used to be we used to look at metrics like time on time on page or total views and things like that. There's now a lot of interest in visibility metrics, right? Being cited or referenced inside these AI answers become increasingly important. So it's a big shift, but I think one that marketers can have some impact on. So when you say visibility, you mean rank becomes less important almost as opposed to share a voice in a lot of ways, right? You wanna make sure that you were cited and referenced a lot more. Share a voice, trust, right? It's a topic you and I have talked already quite a lot about in our first few sessions here, which is you have to build trust, and I'm gonna tell another, I'm gonna share another story here in a minute where that trust was broken very well, and they're paying the price for it. So trust absolutely is critical, that share of voice, that expertise, think is really important. The more that you can be seen as genuinely adding to a conversation, genuinely adding expertise and knowledge that help people make decisions, the better off you're going to be. You know, the things that I would leave CMOs with in this area is, first and foremost, this is just generally good internet hygiene, is respond to negative reviews or comments that are out there, right? It may have been from five years ago, but that stuff is getting pulled into these AI overviews, so go and seek those out and try to respond to them in genuine ways. Don't let your bots answer them because that's going to be very obvious to the user, that's not a real response, but pay attention to those kind of reputational signals. Related to that, I would say is get the old stuff off your website that isn't helping your current story. So you might think that's an old page, we haven't touched that in ten years, we don't offer that service anymore, but nobody ever goes there. That's not the point anymore, the point is Google's and others are going to pull that information in to these overviews to help create that rounded response. So if you have old stuff on your site that isn't relevant to you anymore, or is maybe outdated, get rid of it, because it will start to be a drag on your brand. And the last thing you touched on here just a moment ago is investing in subject matter expert content. Right? Whatever your area of interest and expertise is in your industry, create content that really resonates and helps people answer a question, solve a problem, move forward in a buying decision, and do that with an expert at its... At the ready so that people can feel like this is legitimate. This is not just some AI nonsense that I'm reading. Yeah. I think, we've spoken, or you and me have spoken about this in previous episodes where that element of brand discovery is Yeah. That is just completely changing. I know a lot of research that is out there, which says it's still a tiny proportion. I was attending a a webinar by Rand Fishkin where he was talking about that all AI search desktop... AI desktop search in North America is about 3% or so still, and Google still dominates. But I feel, you know, it's... We can't ignore it because that's what happens. When when you ignore it, you you... That's when you you create Blockbuster and Netflix type stories. Right? Right. So I think for a lot of a lot of brains, you can't ignore the fact that you gotta do things to start showing up in these LLMs and doing it in the correct way. Also, I feel like they are not a new lease in life. It's not that it went away, but I think the importance of making sure that your brand is now referenced out there in positive light in order to increase the authority of your brand holds a whole new level of significance in the world of LLMs. Yeah. Totally. Yeah. Super important. Alright, Aby. Let's move on to your third topic here. You're gonna move away from agents, and you are going to talk about Adobe and a stock impact. What are you gonna tell us? Yes. I thought to keep the theme, would do another agent story, but I am going to give you a break for this Oh, thank you. Goodness. So this one is interesting at a very different level because, you know, the CEO of Adobe, his name is Shantanu Narayen, announced that he plans to step down after eighteen years at the helm. Right? So he's the guy who's really built Adobe to what it is right now. And, yes, Adobe has been struggling. I think year over year, their stock's down, like, 23. In the last couple of years, it's fallen about 20%. But other than that, I mean, he's really been the architect of Adobe's whole genesis of AI and architecture towards being AI positive, and they've dominated the creative field and experience field, you know, in that sense. But I think this is coming at a really important time, for Adobe because it's moved into a cloud model and... Which is super successful, cloud subscription model, and they've seen major long term growth. However, with the rising... With rising AI competition, so you have Canva and Figma who are rolling out generative AI tools and features to let anybody create professional looking content with simple prompts, that has really challenged Adobe's dominance, and Adobe sort of found itself more in the enterprise space as opposed to some of these other tools, which is, you know, more accessible and easy... Easier, you know, shall we say? Can I interject something here? Because this will help even tell your story even better. There was a story yesterday that Figma's stocks fell 12% just in the last few days because Google released their own vibe coding product. And so yet again, you're seeing these massive stock impacts on these companies as some of these other these other players come in. So I know you're probably just hearing that data for the first time, but, like, to me, that continues to tell the story that you're sharing here. So what do you think about the Figma stock drop in addition to the Adobe stock drop? Absolutely. I mean, to me, I was gonna reference it, but the takeaway is really that for all these companies, the quote unquote moat is shrinking. Right? Like, where Adobe has had, you know, a strong moat for years, and it looks with Photoshop and Illustrator. They weren't sort of the de facto standards, but not anymore. Right? So that mode is shrinking. Generative AI or AI in general is sort of a great equalizer where you have suddenly all of these tools and technologies. We just did a session for Nano Banana as part of the Marketing AI Pulse webinar yesterday, and and it's just amazing to see what you can do by prompting. You don't need to be a creative person. You don't need a creative eye. I've seen a lot of horrible things coming out of AI and you can spot it, but obviously the process of production is much, much easier, you know, as opposed to the early days when you had, you know, Photoshop and Illustrator when you needed some level of expertise. So that threshold is coming down. And then on top of that, like the the... For marketers, what is really happening is it's the way you you should really think about it is is changing the way, your martech strategy is going to play out right where now you wanna make sure that if you're tied to... You're not tied too much to a specific ecosystem. Because if you were tied to an Adobe ecosystem and suddenly you now have Canva coming up with all these capabilities, or if you're tied to Canva and Figma's coming up with all of these capabilities, Or if you're tied to Figma, Google's coming up with all of these capabilities. So it really makes... There is no there is no kinda Martech concept or Martech loyalty anymore because of AI is what AI is enabling for marketers to do. So that lowering of those barriers of entry and, you know, speed is becoming important. So you have precision, and, yes, you still need precision in a lot of creative things, in the world of, you know, design and that sort of stuff. But but now with Canva, with templates, with prompting, some of these things are very easily achievable at some crazy speeds, especially derivative content. I know you and me worked on projects where you take one masterpiece of image, and if you want then, you know, create hundreds of derivatives for various platforms and those sort of things, it is really easily achievable, and, you know, you don't need a super expensive enterprise scale platform like in Adobe. So I think there's a lot of uncertainty around that. And, also, I found it kinda weird going back to that. Yeah. I find it I find it weird that they didn't really announce a new CEO or when his... When he's getting off the hot seat, so to speak, because he... They say, hey. They're just announcing he's leaving, and, you know, he's gonna stay around till the time they announce a new successor, which they have not announced. So the timing is unknown. So, yes, it was a stock story, but to me, the implications from a marketing standpoint is clear that the, you know, like I said, shrinking mode, don't... You know, the the speed is becoming super important, from a Martech standpoint, and the concept of Martech loyalty needs to be reexamined. I'm not saying go buy a new Martech tool every week, but to me, something that within your team need to start being a little bit more agile to, you know, accommodate for that, because, again, this world is moving at lightning speed. Yeah. No question. Yep. So, Matt, earlier you referenced trust, and we always, you and me, seem to end up talking about just a lot. I know we... That's near and dear to us, but Grammarly did some things which I did not expect Grammarly to do. Can you talk a little bit about that and tell us what's going on there? Yeah, sure. So Grammarly stepped in it, right? Grammarly is, I think of Grammarly as like the gateway drug to AI for a lot of people, right? It was the first thing that a lot of marketers started to use years ago, right? How to use Grammarly to make your emails better or shorten copy that you've written or change the tone from professional to casual, all those kind of things really started with Grammarly. Grammarly has been a very popular, very well regarded technology. It's been there for a while, and I would say interestingly, Figma stock, as we mentioned a moment ago, Figma stock is dropping because of a new product by Google, that to me tells me that some of these larger companies are starting to release capabilities that are gonna replace some of these niche companies, and I think Grammarly in that bucket is this is a company that does one thing and it does it really well, and they had a very strong reputation, a very committed user group, people who really love the product and would swear by it, but they made what I would call self... Gave themselves what I would call a self inflicted wound, recently. In 2025, they released a new capability within the platform called Expert Review, which just by... On the face of it sounds very cool. It's this idea that Grammarly users within the interface could summon specific experts to review their content, so they could get really specific kinds of feedback. So they could say, know what, I really need an English professor to look at my essay, but it wasn't just English professor generically, it was a specific human being, that was being called upon. And so they had this panel of expert reviewers. There was only one problem with this was that none of these expert reviewers knew that they were an expert reviewer in Grammarly. None of them had been asked to be a part of the platform. None of them are being compensated to be a part of this expert review panel. And in fact, wait for it Aby, one of them was dead. Oh my So you have a panel of expert reviewers on a very well used platform, a very well regarded platform, who literally had no idea that this was happening. So The Verge, a really good outlet and publication, did a story on this where they identified some specific people who were mentioned as expert reviewers. Those people had no idea until The Verge article that they were expert reviewers on Grammarly. So Grammarly now is finding themselves in this moment of, oh my gosh, we didn't ask any of these people if they would be expert reviewers, we're not compensating them, we're just going to bury it in our terms and conditions that these are not associated with us, but we're just using them as like an expert point of view. So they try to have it both ways of we want to get the value of the voice of these people and the talent of these people and the reputation of these people, but we're not gonna actually involve them in any way, and they're gonna get completely caught off guard by this. So it was a totally bizarre story, and to where you started, a really easy and fast way to break a decade's worth of trust building. I'm I'm just wondering for a brand like Grammarly, I mean, they're pretty well respected. Did they just... Was this an oversight or was this something that it was a a feature that they launched without really thinking about the consequences, or was it not an adequate quality control? I mean, obviously, we don't know, but I'm just wondering how did this even happen? I don't know exactly how or why this happened, but what it feels like from what I've seen and read is that they really got caught up in the ship first, ask questions later ethos in AI. Right? It's like everybody's trying to out ship each other. Here's a new function. Here's a new model. Here's a new capability, whatever it is, everybody's trying to beat each other to the market. And I think that's causing some bad decision making. Right? I think it's causing people to not really think strategically about their product and what that product really is trying to become. And I think it's to me a further indication that marketing leaders, the folks who are listening here today, really need to pay attention to brand and to the relationships that they have with their customers, right? Because it's, we all know this, really hard to build a good reputation, but really easy to lose it. And I think this Grammarly case is a perfect example that this is a product that people would swear by, and all of a sudden they're getting terrible press at a time when I'm sure they cannot afford to get terrible press, because all the other big companies out there are putting capabilities in their platforms that can do a lot of what Grammarly does. So I think you have to be very careful and cautious about what you're building into your products, how you're positioning your products, how you're adopting AI, how you're talking about adopting AI. And I think even before that is start with a like, should we even use AI here filter? Right? I think a lot of folks just are naturally responding to let's just do AI rather than does it make sense to do it in this specific case? What are the pros and cons? And I think the more that people can take that deep breath and say, does this make sense? Are we going to get ourselves in trouble? Have we thought through the possible consequences of this? Not that I want people to get frozen in indecision, but I think this to me was, this is an example of a completely unnecessary hit to their brand. It's not like this is core functionality, it's not like this is how they go to market, this is just a nice to have feature that probably not that many people really cared an awful lot about, and all of a sudden it's had a massive impact on their brand. Yeah. I mean, you make a good point for a lot of CMOs who are their marketing departments, some marketing products that have AI in it. There is intense pressure for making sure that you're talking about AI, like AI washing AKA. Right? Now Yep. It's all made up or whatever, but to me, certainly, I'm a marketer. I know how I can recognize spin a little bit, but the... Trying to understand what should be the guardrails in place. Because, again, like you said, the consequences, obviously, don't know the, you know, circumstances under which this Grammarly situation happened. But to all CMOs, what would your words of advice be? Dave, a few different things. One is, I mentioned it a moment ago, but does it make sense for us to be developing this functionality or sharing this functionality? Does this need to be AI? Yes, there's pressure to have the latest AI, but to your point, a lot of that is becoming noise, and if that noise becomes something that impacts your brand like this, man, that's not worth it. So really take the time, I think, to be strategic and thoughtful about should we even do this with AI or not? I think really clear guardrails for AI and customer facing experiences is important back to the conversations that we've had today multiple times around governance, right, define when and how AI is allowed to show up and it's not allowed to show up. You and I did a project last year where we went through a very deep dive with a large complex company to help really figure out where and how they could use AI in their business, in their marketing and not. It's an important conversation to really have and not just assume that it's gonna work itself out. And then last but not least, and this is marketing 101, but talk to your customers before you roll new stuff out. Understand what are they looking for, will this be seen as valuable or not, will it even matter to them if you add this feature and functionality? We all want to build to vision, but I think a lot of these companies also need to build to what their clients really want and need, and I think finding that balance is super important. So go out, talk to your folks, and hear what they want from you, and use that as at least one of your decision making points as you go forward. The world is changing, right? The way that brands need to appear is changing, and the way that trust is built and broken is changing. And so I would just say, don't do what Grammarly did here, is burn a decade's worth of trust to the ground for a nice to have feature. Great advice. And, Matt, I think that is a wrap. That was our last story. We have been talking for almost an hour. Yeah. I think this was awesome. Thank you very much. And to our listeners, make sure to connect up with Matt and me on LinkedIn. And if you have any questions or suggestions for any of our future episodes or you have any burning questions on how to use AI or leverage AI in the world of marketing, throw your ideas in. You know, DM Matt or me, and we can cover that in one of the future episodes. So, Matt, thank you very much, and, have a good one. You too. Thanks. See you soon. Thanks for listening. Big thanks to Spark Novus and the Marketing AI Pulse community. You can join the community at marketingaipulse.com.
