This Week in Marketing Technology, AI, and CX Podcasts | July 23, 2026
Every conversation this week circles the same uncomfortable finding: the constraint on marketing performance is no longer capability. It is the distance between what an organization knows and what it can actually act on. Stephanie Trunzo, Chief Executive Officer at MERGE, argues that artificial intelligence (AI) is not creating organizational gaps so much as revealing ones that were already there, and that the work of closing them is cultural rather than technical. Tal Peretz, CEO and Co-Founder of Onfire, makes the commercial version of the same case: once every revenue team can generate personalized outreach at volume, output stops differentiating anyone and the advantage moves to whoever holds the right data and context. Lou Carbone, Professor of Practice at Michigan State University and Founder and Chief Experience Officer at Experience Engineering, pushes the argument further back still, into the unconscious signals that shape how customers actually feel about a brand — territory that process maps and satisfaction scores never reach. Alexi Hatch, Chief Marketing Officer at Acoustic, names the operational cost of that distance directly, describing marketing’s central problem as one of timing rather than data. And Sidd Motwani and Di Lyngholm of Malachyte show what happens when that lag is engineered out of an ecommerce merchandising system entirely. Our bonus pick from The Artificial Intelligence Show supplies the survey evidence underneath all five: most employees are now using AI, and most of them have been told nothing about what to do with the time it gives back.
MERGE CEO Stephanie Trunzo on keeping essential human elements during your AI transformation
Stephanie Trunzo, Chief Executive Officer at MERGE — a marketing and technology agency where roughly 30% of the business is consumer-native brands including American Express and Subway and roughly 70% sits in health and wellness, including companies like Oura Ring — frames AI adoption as a paradigm shift rather than a tooling decision, and offers a diagnostic she uses with clients: the gap between expectation and reality. When expectation outruns reality you operate in deficit, and when reality outruns expectation you operate in abundance, but the discipline is being honest about which side of that equation you actually need to adjust. Her second framework is a pendulum, where the point of control is not the technology but your value proposition and differentiators; teams getting whipsawed between metaverse-style overinvestment and reflexive caution usually lack a true north, and when a leader says “we need an AI strategy,” Trunzo hears that they need a strategy first. She describes AI as producing smooth surfaces while humans supply texture — the half-baked thought, the unpolished draft, the em dash, the deliberate typo left in an email as evidence a person was there — and argues that texture is what other humans grip onto, which matters enormously in regulated categories where health and financial information make emotional connection non-negotiable. On measurement she is blunt that nobody yet knows the new key performance indicators (KPIs), comparing today’s borrowed metrics to horsepower, a psychological bridge from horses to steam engines that measured nothing useful, and she insists that the return on an AI investment is a return against organizational change, workflow redesign, and strategy work rather than a technology ROI on a tools decision. MERGE runs an internal AI Garage that serves as both a secure sandbox for enablement and a build environment for agents and platforms, inward and outward facing at once. Her closing charge is a sequencing one: pull answer engine optimization (AEO), generative engine optimization (GEO), and similar concerns earlier in the creation cycle rather than retrofitting them, and never outsource creativity, ideas, or critical judgment.
Onfire CEO Tal Peretz on AI investments that actually drive revenue
Tal Peretz, CEO and Co-Founder of Onfire and previously Chief Technology Officer at the identity security startup OwnID, makes an argument that runs against the prevailing optimism: generic AI has made go-to-market harder rather than easier, because once every team can produce mass personalization at scale the bar simply rises and volume stops being a differentiator. His alternative is vertical AI, meaning systems specialized to a category rather than general-purpose ones, and Onfire’s chosen vertical is companies selling to technical buyers — the chief technology officers, chief information security officers, and chief information officers who evaluate tools in Reddit threads, Slack groups, and Discord servers rather than on vendor websites. The limitation of first-party data, in his framing, is structural rather than a quality problem: your calls and meetings capture only what a prospect chose to tell you, while competitor comparisons, internal success plans, and buying-committee members you have never met get discussed in rooms you are not in. Turning those rooms into pipeline takes three distinct technical layers, and Peretz breaks each one out — locating communities that in some cases are not indexed on the public web and had to be found manually, classifying messages through Onfire’s own models built on public large language models plus proprietary data to separate genuine evaluation and pricing questions from ordinary conversation, and then resolving pseudonymous accounts to real individuals through the public footprints those users leave across communities. He walks through a public cybersecurity company running a data-first strategy where every external source lands in a Snowflake data lake alongside Salesforce records and call-recording data, with account executives able to prompt an AI agent directly against that store. On measurement he sequences leading indicators down the funnel — relevant prospects discovered, prospects engaged, meetings booked, opportunities created — beneath a single top-line metric of revenue per rep. And his first-step advice avoids the transformation program entirely: name one owner inside revenue or marketing operations with time to think from first principles, decide where the data will live, then take one workflow at a time, since roughly 80% of a seller’s time currently goes to manual work that AI is genuinely good at absorbing.
Lou Carbone on a new understanding of brand
Lou Carbone, Professor of Practice at Michigan State University and Founder and Chief Experience Officer at Experience Engineering, has been called the godfather of experience management, authored Clued In, and developed the proprietary Experience Engineering methodology; in this conversation he calls for what he terms a reformation in how brand is understood. His starting claim is that the industrial-age assumption that companies control their brands is simply wrong — consumers do — and that the currency of experience is memory, which means the discipline’s inherited habit of eliminating defects rather than creating distinctive value guarantees commoditization. What replaces it is brand ethos, the feeling that an organization is an institution rather than a company, and the operational test he offers is whether a customer would mourn the loss of your business if it disappeared tomorrow. Getting there requires traveling what he calls the maze of the customer’s mind to understand how customers think rather than what they think, then embedding experience clues — humanic, mechanic, and functional — that create emotional resonance, and measuring emotional congruency rather than a score. Carbone is scathing about the score itself, arguing that the pressure on customer experience teams to prove ROI has actively hurt the profession because a number becomes justification rather than results, and he dismisses net promoter score’s central question by pointing out that as a VP of marketing at National Car decades ago he already knew he would rather measure whether someone actually referred a customer than whether they said they would. His practical illustration is a barbershop, Truefitt & Hill, that gave him a poor first haircut and kept his business for years because every clue in the visit — the tea, the shoeshine, the shave — was engineered around how he would feel. On AI he is equally direct: it is not a bolt-on, and treating it that way misunderstands it. AI belongs at the center of an operating system that is fundamentally human, the fusion of understanding human dynamics with the technology itself. His most quoted line remains a double negative that functions as a strategic warning — you cannot not have an experience; the only question is how managed or how haphazard it is.
Acoustic CMO Alexi Hatch on rebuilding your customer relationship strategy
Alexi Hatch, Chief Marketing Officer at Acoustic, where she leads global marketing, brand, communications, product marketing, demand generation, and go-to-market strategy, opens with a claim about intent that reframes the entire third-party data debate: intent has a half-life and begins decaying the moment it appears, so by the time a signal has been consented to, collected, packaged, put in a database, purchased, and activated, the buyer has usually already decided. That leads to her sharpest reframe of the conversation — marketers do not have a data problem, they have a timing problem — and to a first-party data standard that has less to do with volume than with completeness, tagging, and how fast information can move from one system to the next. On measurement she is willing to pick a fight, saying every problem is an attribution problem and that even a perfect multi-touch model would be useless if you could not replicate what it revealed; she would rather decide today on 85% of the data than wait three weeks and lose the window, and she argues the industry’s pursuit of the perfect model has actively slowed marketers down. The KPIs she wants in its place shift from net-new contacts added to how live a database actually is: how many customers return, what new information is being collected on them, how quickly the data refreshes, and how much revenue comes from existing customers rather than the quarter’s total. Her distinction between personalized content and personalized experience is the one most worth carrying into a planning meeting — AI has commoditized content production to the point that any marketer can do what was difficult a year ago, but content is not activation, which means the competitive advantage sits in the experience and in the stack that can deliver it in the moment. Looking a year out, she expects agents to be making purchases at something closer to mainstream scale, raising the question of how you optimize for a non-human audience, and she names the lag between signal and action as the last remaining competitive moat.
One Amazing Thing About Malachyte with Sidd Motwani and Di Lyngholm
Sidd Motwani, Co-Founder and CEO of Malachyte, and Di Lyngholm, VP of Product at Malachyte, take on the problem every ecommerce merchandiser recognizes and few solve — how to adapt search results, product carousels, and the rest of the on-site experience continuously, including from a cold start when there is no history to draw on. Malachyte’s position is that retailers should replace their stack of disconnected personalization tools with a single real-time intelligence engine, moving the industry off static rules and periodic experimentation and onto autonomous decision-making that happens in the moment. Motwani built personalization systems at Spotify and at Booking.com and Priceline before founding the company, and the conviction he carried out of that work sharpens the week’s theme considerably: the future of commerce lies not in collecting more data but in intelligence that learns, adapts, and decides as customers move. It is the shortest conversation in this week’s set and the most concrete demonstration of what Hatch means by closing the gap between signal and action.
Bonus Pick: [The AI Show Episode 225]: GPT-5.6, ChatGPT Work, Enterprise Agents, AI 2040 & Apple Sues OpenAI — The Artificial Intelligence Show
Paul Roetzer, founder and CEO of SmarterX and Marketing AI Institute, and Mike Kaput, Chief Content Officer at SmarterX, spend the middle of this episode on the enterprise reality underneath the model releases, and that segment is what makes it the right companion to this week’s Agile Brand conversations. The Boston Consulting Group’s fourth annual AI at Work survey of nearly 12,000 employees across a dozen global markets found that 74% of frontline workers are now regular AI users, up 23 points year over year, and that 42% of AI users save the equivalent of a full day of work each week — with time savings running higher in marketing at 60%. The finding that matters, though, is what happens next: 66% receive little or no guidance on what to do with the time they save, more than half say they are not reinvesting it in strategic work, only a third find leadership’s AI communication clear, and just 28% see a strong connection between what leaders say and what their organizations actually do. Agent adoption has more than doubled to 30% from 13%, and 61% of employees believe agents could handle at least half their job within three years, while governance has not kept pace. Roetzer and Kaput also work through Aaron Levie’s argument that agents force an operating-model problem because the most valuable processes cut across the silos companies are organized into, and Satya Nadella’s essay on what he calls the reverse information paradox — the idea that using purchased intelligence requires surrendering the proprietary knowledge that made you distinctive. Set beside Trunzo on change management as the real ROI question and Peretz on naming a single owner before launching a program, the survey data reads less like news than like confirmation.
Taken together, these five conversations describe a discipline that has solved for capability and not yet for consequence. The tools can generate the content, classify the signal, and personalize the carousel; what remains scarce is the strategic clarity to decide which of those outputs deserves to reach a customer, and the organizational speed to act while the moment is still open. Trunzo, Peretz, Carbone, Hatch, Motwani, and Lyngholm arrive at that conclusion from five different starting points, which is usually a sign it is worth taking seriously. See you next week!
Frequently Asked Questions
What are the best marketing, AI, and CX podcast episodes for the week of July 23, 2026? This week’s standouts are Stephanie Trunzo of MERGE on keeping human creativity and judgment intact during AI transformation, Tal Peretz of Onfire on vertical AI for go-to-market teams, Lou Carbone of Experience Engineering on a new understanding of brand, Alexi Hatch of Acoustic on rebuilding customer relationship strategy, and Sidd Motwani and Di Lyngholm of Malachyte on real-time ecommerce personalization. All five appear on Greg Kihlström’s shows, The Agile Brand with Greg Kihlström and One Amazing Thing with Greg Kihlström. Episode 225 of The Artificial Intelligence Show with Paul Roetzer and Mike Kaput rounds out the week with survey data on enterprise AI adoption.
What does MERGE CEO Stephanie Trunzo say about measuring the ROI of AI investments? Stephanie Trunzo, Chief Executive Officer at MERGE, argues that the return on an AI investment is not a technology ROI on a tools decision but a return against all the organizational change that has to accompany it, including workflow redesign and strategy work. Stephanie Trunzo compares today’s borrowed KPIs to horsepower, a metric that existed to bridge people’s understanding from horses to steam engines rather than to measure anything useful. Her practical recommendation is to define a hypothesis, scope small projects outside the core business, and learn what is worth measuring by doing the work.
How does Onfire CEO Tal Peretz define vertical AI for go-to-market teams? Tal Peretz, CEO and Co-Founder of Onfire, defines vertical AI as a system specialized to a specific industry or buyer category rather than a general-purpose tool, on the reasoning that when general AI capability is available to everyone a horizontal tool gives every team the same lift and therefore no advantage. Onfire’s vertical is companies selling to technical buyers such as chief technology officers, chief information security officers, and chief information officers. Tal Peretz identifies three technical layers required to make it work: finding communities that are sometimes not indexed on the public web, classifying which messages signal genuine buying intent, and resolving pseudonymous accounts to real individuals.
What is Lou Carbone’s argument about brand and customer experience measurement? Lou Carbone, Professor of Practice at Michigan State University and Founder and Chief Experience Officer at Experience Engineering, argues that a brand is the cumulative effect of every customer interaction rather than what a company says about itself, and that the currency of experience is memory. Lou Carbone believes the pressure to prove ROI has hurt the customer experience profession by turning scores into justification rather than results, and he criticizes net promoter score for asking whether someone would recommend a business instead of measuring whether they actually did. His alternative is to measure emotional congruency and to manage what he calls experience clues — humanic, mechanic, and functional.
Why does Acoustic CMO Alexi Hatch say marketers have a timing problem rather than a data problem? Alexi Hatch, Chief Marketing Officer at Acoustic, argues that buyer intent has a half-life and starts decaying the moment it appears, so the delay between a signal being generated and a brand acting on it destroys most of its value. Alexi Hatch says marketers already have an enormous amount of data and that the real constraint is how quickly it moves between systems and reaches the customer. She would rather make a decision today on 85% of the data than wait three weeks for a complete picture, and she calls the lag between signal and action the last remaining competitive moat.
What does Malachyte do for ecommerce personalization? Malachyte, co-founded by CEO Sidd Motwani with Di Lyngholm serving as VP of Product, is an AI technology company that replaces a retailer’s disconnected personalization tools with a single real-time intelligence engine. The platform continuously adapts elements like search results and product carousels for individual customers, including from a cold start with no prior history. Sidd Motwani built personalization systems at Spotify and at Booking.com and Priceline before founding Malachyte, and argues that the future of commerce lies in intelligence that learns and decides in the moment rather than in collecting more data.
What is the common theme across this week’s marketing, AI, and CX podcast episodes? The through-line is that the constraint on marketing performance has shifted from capability to consequence — organizations can now generate content, classify signals, and personalize experiences at scale, but struggle to act on what they know while it still matters. Stephanie Trunzo of MERGE frames it as a cultural and strategic gap, Tal Peretz of Onfire as a data and context gap, Lou Carbone of Experience Engineering as a gap between measurement and meaning, and Alexi Hatch of Acoustic as a timing gap. Boston Consulting Group survey data cited on The Artificial Intelligence Show supports the pattern: 42% of AI users save a full day of work per week, but 66% receive little or no guidance on what to do with that time.
