Tom Wentworth on Taste, CMS Evolution & the Future of AI-Driven Marketing

ICYMI: Tom Wentworth Explains How Taste Is the New Competitive Advantage in an AI-Powered World

“Marketing output is going to be AI adoption times taste squared.”

Tom Wentworth, CMO of incident.io, joins 10 Minute Martech host Sara Faatz to unpack a powerful idea: AI will soon level the playing field, but taste will separate the leaders from the laggards.

From the role of design in an AI-saturated world to why attribution still drives him up the wall, Tom delivers an unfiltered, deeply practical view of what really matters as AI becomes embedded in every marketing workflow.

He believes the biggest advantage for marketers isn’t more automation, it’s the ability to craft experiences that feel intentional, human and high taste. He also explains why CMS platforms will become even more relevant in the AI era, how LLM discoverability works, and why the best minds in AI aren’t on LinkedIn … but on X.

If you care about content quality, creative differentiation, CMS strategy or building modern AI-first workflows, this episode is packed with insights that cut straight to the point.

Tom’s Big Idea: Taste Is the New Differentiator

Tom believes we’re in a temporary arbitrage window where AI can still give brands a competitive edge, but that window is closing fast. Soon, everyone will have access to:

• The same tools
• The same models
• The same data
• The same automation

What won’t be identical? Taste.

The human ability to choose, refine, elevate and say: “No! This is what great looks like.”

“At some point, we all narrow in on the same tooling,” Tom says. “Taste is what creates winners and losers.”

In Tom’s world, design agencies of the future will use AI to build faster but will win based on their ability to deliver high-taste experiences, not commodity templates.

Tom’s 10 Most Memorable Moments on 10 Minute Martech

1. Taste > Tools

AI accelerates everything except good judgment. Tom argues that as AI democratizes production, taste becomes the exponential differentiator.

2. Creativity Isn’t Getting Replaced

LLMs are built on averages. Great designers aren’t. As long as models are trained on collective output, creative excellence will remain human-led.

3. The CMS Is Becoming More Important

AI eliminates the hardest part of CMS projects, the site build. What used to take 12 months can now be prototyped in hours. That makes CMS platforms more valuable, not less.

4. AI Unlocks Faster, Better Personalization

With the build process streamlined, marketers can finally launch personalized experiences at the speed users expect. And yes, taste matters here, too.

5. LLM Discoverability Is the New SEO

To rank in AI responses, brands must understand:

• Which LLMs matter (ChatGPT, Gemini, Claude, Grok, Perplexity)
• What content they cite
• Where those sources originate (often Reddit!)
  SEO isn’t dead but it’s evolving fast.

6. Reddit Is a Dominant LLM Source

If you’re not present on Reddit, you’re invisible in many AI-generated answers. Content strategy now includes community strategy.

7. Attribution Is ‘the Dumbest Idea in the World’

Tom’s words and he stands by them. Today’s customer journeys are far too complex for any single-touch or even multi-touch model to capture. He refuses to work somewhere with outdated attribution expectations.

8. Rethinking the MQL

The classic MQL formula, fit + arbitrary activity, is now broken. At incident.io, MQLs = people who expressly want to talk to sales. Everything else is a “gry signal, supported by AI-powered insights."

9. X Is Where the Real AI Innovation Lives

LinkedIn is full of hot takes and vendor agendas. On X, Tom finds the builders, the engineers, hackers and founders experimenting in real time. His recommendation: follow Dharmesh Shah for genuinely useful AI learnings.

10. MCP Is the Most Interesting Development in AI Right Now

Tom shares a real example:
His email client (Shortwave) uses Model Context Protocol to coordinate email, tasks, transcripts, meetings, Slack and web research, all through one prompt. It’s how he plans his entire day. This is the future of agentic workflows.

Tom’s Martech Hot Take

Attribution is fundamentally broken, and holding marketers to a flawed attribution model is a dealbreaker.

Journeys are too complex, channels too fragmented and AI too interwoven for any single model to accurately assign credit. It’s time to stop forcing a simple answer onto a complex reality.

Who Tom Follows

Tom’s go-to sources for real, unpolished, in-the-trenches AI insights include:

• Dharmesh Shah (HubSpot founder, prolific AI experimenter)
• The builders and engineers posting cutting-edge work on X
• Deep-dive posts on SaaStr and technical AI communities

Listen to the 10 Minute Martech Episode Now

• Watch the full episode on YouTube

• Listen on Spotify

• Find it on Apple Podcasts

Next Up in the 10 Minute Martech ICYMI Series!

Reed Hansen, Chief Growth Officer at MarketSurge, joins Sara to unpack how the marketer’s role is being redefined in the age of AI.

Full Transcript: Tom Wentworth

Here’s the full transcript to keep you transfixed. Every insight, every quote, unedited and unforgettable.

Sara Faatz: I’m Sara Faatz. I lead community and awareness at Progress, and this is 10 Minute Martech.

Tom Wentworth: If I’m a design agency today, I use the tooling now to speed up building, but I focus all my effort on taste. I tell my clients that I’m going to help you deploy a high-taste experience, so that… And I’m not putting all my budget toward the stuff that now can be automated with AI.

Sara Faatz: That’s Tom Wentworth, Chief Marketing Officer at incident.io. Let’s get started. Tom, thank you so much for joining us today. I want to start by asking you, what is the one thing that’s keeping you up at night right now?

Tom Wentworth: So everybody talks about AI as the thing that’s changing marketing, and I think there’s a moment right now where, as a company, as a brand, you can use AI in a way that can differentiate your business, but I think that’s a short-term arbitrage opportunity. I’ve now landed that the thing that’s going to separate marketing going forward is going to be taste, which is hard for me because I’m a math nerd. I don’t have taste, but I put this formula… I’m a math guy, I put a formula together, and the formula was like marketing output is going to be AI adoption times taste squared, meaning that taste is going to scale exponentially further than AI adoption is going to. At some point, we’re all going to narrow in on the same tooling, the same sets of data, the same models, but taste is going to be what creates winners and losers. I think B2C companies have proven this is how they’ve succeeded. I think we’re going to get back to that. Things like language, the human touch, right? Keeping the humanity in marketing sounds like, why would you do that, Tom? AI is taking over the world, but I think it is honestly like what is going to create leverage, like even right now. So short term, I think AI is a huge arbitrage opportunity. Long term, it’s all about taste.

Sara Faatz: So is it safe to say you don’t see design being eaten up by AI, that that creativity remains within the human?

Tom Wentworth: This is the whole thing about AI is models have to be trained on something, right? So if models are trained on the collective, and essentially LLMs steal stuff, right? So if they’re stealing the collective average of every image on the internet, they’re going to get average output. I would bet on really great design. Maybe someday this changes, who knows? But with the state of the art right now, a really great designer is able to do things that no LLM can even come close to.

Sara Faatz: Let’s talk about the CMS space a little bit. You’ve been around that for a very long time. Tell me, where do you think the CMS lives moving forward when it comes to AI being in the mix?

Tom Wentworth: It takes away potentially the biggest challenge of deploying a CMS, which is never about the CMS itself. Deploying a CMS, training up users, creating content types, defining all the things that you need to organizationally was never the hard part. The hard part was a site build. And now with AI, I can prototype front ends in hours, not months, days, years. The average CMS project used to take a year, which is insane.

Sara Faatz: Right, yeah. Does that change the role of the CMS in your opinion, or does it make it more relevant?

Tom Wentworth: It makes it… In my mind, it makes more relevant because now we finally will be able to move faster. Marketers will be able to get more personalized experiences out to people faster. We’ll be able to build these bespoke design experiences. I talked about the taste that we talked about now. If I’m a design agency today, I use the tooling now to speed up building, but I focus all my effort on taste. I tell my clients that I’m going to help you deploy a high taste experience so that I’m not… And I’m not putting all my budget towards the stuff that now can be automated with AI.

Sara Faatz: Right. LLM discoverability and a CMS should go hand in hand, right? What should you be thinking about to make sure that you are discoverable? And how does your CMS play a role in that from an LLM perspective?

Tom Wentworth: You have to decide the kinds of things you want to be found for. What you do is you say, I want to rank for best CMS, a simple one. There are probably 100 different ways that you could write that into an LLM. There are four or five different LLMs that you care about, Gemini, Perplexity, ChatGPT, maybe Grok, certainly Claude. You want to understand now, across all those LLMs and across the 100 permutations of the thing, how are you doing? What content is being cited right now? Where are the LLMs pulling their stuff from? They didn’t just create it out of thin air. And then your content strategy is ultimately about either addressing the gaps or going to the source material, which often is places like Reddit. Reddit is one of the most dominant sources of LLM information. So if you’re not on Reddit, if you’re not out there actively thinking about that, you’re going to fall behind pretty quickly. But it changes the whole conversation. It’s not all too dissimilar from SEO, but it’s different enough that it requires, I think, a slightly different skill set, and the tooling will certainly be different.

Sara Faatz: Exactly. Well, and that goes back to what you talked about before, which is that human interaction, right? People want to…

Tom Wentworth: Totally.

Sara Faatz: They want to engage with real humans. Tell me what some of your current hot takes are.

Tom Wentworth: Attribution is always my hot take. It’s the dumbest idea in the world that we think that we can take what is a very complex journey into like one field we dump into Salesforce. It was inbound. No, it was outbound. No, our partner brought this deal to us. Like it is insanely difficult to actually understand. It’s like somebody started with ChatGPT, they landed on your demo site, they went over to Reddit to do some research, and then they search your brand and converted. Like, what do you give that credit to? Right now it’s actually much more complicated than that in the real world. We try to look at attribution like we sell products on Instagram, where it’s like we’re at e-commerce, like you put an ad up, somebody clicked on it, and they bought. That is just not how the world works, at least in my universe. I would never work at a company where I was held to some sort of archaic attribution model. Like I would either… If I would just not go work at the company, I’d quit. It’s that strongly of a belief for me. This is one of those non-negotiables. I don’t know if that’s even a hot take anymore, but that’s certainly my biggest pet peeve.

Sara Faatz: Do you think that the MQL is still a valid measurement, should we be looking…

Tom Wentworth: I think MQL… So it is, yeah, it’s a valid measurement, but not in the way that Martech vendors for the past 15 years have told us it was. It was essentially, we have some basic fit attributes, like do they have the right title, and are they at a company, and then we have some activity attributes. How many webinars did they go to? How many emails did they open? Those were not statistically significant ways to understand something. My current company, our MQLs are very simply people that expressly want to talk to us. Like they’ve said either through just, I want to talk to sales, or they start in a trial, but the rest of the world is gray signals. I think of it as sort of warm outbound. My job is to activate people and get them engaging with us in lots of different ways, and I think it’s then the seller’s job on my team to be able to reach out to people at the right time with the right message based on the understanding of the things they’re doing. So if somebody’s on a high intent page, that’s going to be a different conversation than if somebody, we talked to somebody at an event, and I equip the sellers with all the signals that could make sense, all the messages that might make sense for all different personas. This is where AI is great because you can scale in a way you could never do before. If I have 10 signals and 10 personas, that creates a pretty significant complexity map, but AI is perfect for that. That’s where AI shines.

Sara Faatz: Right. Awesome. So tell me, where do you go right now for either inspiration or information?

Tom Wentworth: So speaking of hot takes, LinkedIn is a dumpster buyer of hot takes on AI. I saw a post, last night I was flying back from London. It was like 11 o’clock Eastern time. So whatever time that was in London, I was still awake because I was all wired from the trip home. I saw a post from Saster talking about real insights in the trenches from AI, like what’s working today. That was super cool. I’m going to watch it again and take some actions away because either on LinkedIn, you either have somebody who has never done the thing before they’re talking about or it’s a vendor who’s pushing their own view of the world on you. It’s really hard to find people that are doing it. Honestly, this is going to sound like the most egotistical, narcissistic thing ever, but you should follow me, because I’m out there trying to do the things and I’m sharing the stuff that works and that doesn’t work. And I can promise you I have no agenda. I just like sharing what I think is interesting stuff. So I just, all of a sudden, I turned this into a commercial for me.

Sara Faatz: Well, we’ll make sure it’s in there anyway, whether we want it…

Tom Wentworth: Honestly, so like, here’s the thing, the best source for this and I don’t want to wade into this one. This will also get me in trouble. But all the people that are really doing interesting stuff on AI, not at the Martech GTM level, but the actual people doing it are on X. Believe what you want about the X platform. I don’t care, but that is where the actual people doing really interesting cutting-edge stuff on AI actually live. And it’s not like Sam Altman level. It’s people trying cool stuff. The CEOs of all the LLM companies have this big vision that is not the reality of the world that we live in today. We’re not… AGI is not imminent. The machines are not going to take our jobs in six months, but there are cool things being done with AI every day. And I learn a lot there. I learn a lot. Here’s one specific recommendation. Dharmesh Shah, who has to be a multi-billionaire founder of HubSpot. That dude is grinding on AI every single night. I don’t know why he works. He literally is insanely wealthy, but he’s out there doing cool stuff every single day. He’s great to follow. He’s on LinkedIn. He shares real learnings. Good person like that. And one very specific recommendation, if you don’t know MCP, Model Control Protocol, or play with it, it is the most interesting thing to happen in the AI world in the past couple of years. Definitely check out what you can do with MCP. Can I share a real example with you?

Sara Faatz: Yeah, absolutely.

Tom Wentworth: I talk a lot. I don’t know if you have time for this, but my real example is my email client supports MCP. It’s called Shortwave. It’s great. And it is connected to Linear, which we use for tasks, like task management system. Obviously, it’s connected to my email and my calendar. It’s connected to all my call transcripts, and it’s collected to Slack. So every day in the morning when I wake up, I don’t check my email. I run a prompt that I saved called My Day, and I say, look at everything holistically and tell me what’s important. Go do research, because this thing can also access the web. So I’m meeting with somebody, go research them, and give me a nice summarized list of my entire day. And then after that, I can say things like, block off my calendar so I can get the work done that you told me was important. Or I can say, cancel the meeting with this person, because it knows how to do all of these things with MCP. And it goes across all of it. It correlates it. And it will be able to say things like, “Hey, you have this linear task, and you just got an email about it. Would you like me to update the linear task with the information you got in your email so you stay in sync?” It’s crazy.

Sara Faatz: Well, Tom, this has been amazing. Thank you so much for your time, and I really appreciate all of your insight.

Tom Wentworth: Would you want any more narcissism, or are we done with narcissism today?

Sara Faatz: If you have more, bring it on.

Tom Wentworth: No, we’re done. We had enough. Obviously, we had an overabundance of it. Thank you so much. It was really fun.

Sara Faatz: Hey, listeners, thanks for tuning in. Make sure you like and subscribe wherever you get your podcasts. Until next time, I’m Sara Faatz, and this is 10 Minute Martech.