There is no shortage of executives willing to talk about responsible AI. There is a considerably shorter list of those willing to be honest about what it actually costs — and what it demands of the people whose names are on the org chart when things go wrong.

Peter Yeung, Chief Information Officer at Optimizely, is in the second group. A former practicing attorney with 18 years at the bar before moving into technology leadership, he brings an unusual combination of legal precision and operational candor to questions that the industry too often answers with carefully worded reassurance.

In a wide-ranging conversation, Yeung addresses the governance paradox at the heart of enterprise AI — how to move quickly without moving recklessly — and argues that accountability, far from being a legal fiction, is a structure that leaders must be willing to sign their name to. He also takes on explainability under GDPR, the data minimization debate, and the accelerating fragmentation of the global data landscape.

His answers are not always comfortable. That is precisely what makes them worth reading.

Excerpts from the interview; 

Companies are rolling out AI faster than governance can keep up. Is ‘responsible AI’ just a story businesses tell to move quickly, or do you truly think governance can match the pace of deployment?

The companies actually getting value from AI aren’t treating governance as a brake; they’re building it into how they scale. Most of us started broadly: put the tools in people’s hands, see what sticks. That phase served its purpose, but what’s working now is the opposite — picking a handful of high-impact use cases and making sure the data, controls, and workflows behind them are genuinely solid, secure, and trustworthy. Done right, governance accelerates things by cutting rework, risk, inaccuracies, and fragmentation.

That said, I’d be lying if I said governance doesn’t have a cost. The fastest innovation I’ve seen on AI happens in the messy middle — small teams shipping fast, breaking things, learning in days rather than quarters. The moment you wrap that in review boards, data classifications, and approval workflows, you do slow it down. That’s just the reality. The trick isn’t pretending the trade-off doesn’t exist; it’s finding the right balance for where you are. Too little governance and you end up with a graveyard of pilots and a compliance problem. Too much and you kill the energy that made AI exciting in the first place.

Responsible AI isn’t a layer you bolt on top of performance; it’s what allows AI to graduate from experimentation into something the business can actually rely on. But you have to be honest that getting the balance right is the work.

When AI systems use flawed or unclear data and cause harm, responsibility is often spread among teams and vendors. Right now, isn’t the idea of clear accountability in AI mostly just a legal fiction? 

As CIO at Optimizely, with both the CISO and Trust organization reporting into me, I’d push back on the idea that accountability is a legal fiction — but I understand why people frame it that way. AI accountability is more complex than in traditional systems because it spans multiple teams: the people sourcing the data, the people building or selecting the models, and the people deciding how outputs are actually used in the business. Spread that across vendors, too, and yes, it can feel diffuse. If you then include my statement above, which calls for empowering individuals within the business to innovate at speed, the task becomes daunting. 

But the way I look at it, regardless of the actor — vendor, third-party model, internal team, or individual employee — we are ultimately accountable, both internally and to our customers, for the end result. That accountability can’t be outsourced. The vendor contract doesn’t absolve us. The model provider doesn’t absolve us. If something goes wrong, our customers don’t care about the seven hops in the supply chain; they care that we own it. 

What makes that real, rather than rhetorical, is structure. We treat AI like any other critical business process: explicit ownership of data inputs, clear responsibility for model deployment, and a named, accountable owner for outcomes in production. Without that, accountability genuinely does dilute across vendors and teams, and that’s where the “legal fiction” critique starts to land. With it, you create a clear line of responsibility even in a distributed system, and you give the CISO and Trust functions something concrete to govern against. 

So it’s not a fiction. It’s just harder, and it requires leaders to actually sign their names.

Rules like GDPR require that automated decisions be explainable. But big AI systems often cannot give real reasons for their choices. Are we trying to enforce laws that no longer fit the world we live in?

Having practiced as an attorney for 18 years, I’d say the question is sharper than the framing suggests — but the answer isn’t quite “the laws no longer fit.” It’s that the laws were never as clear as people assume.

GDPR’s intent is absolutely still relevant: to protect individuals and hold companies accountable for automated decisions that affect them. That hasn’t aged. But read Article 22 alongside Articles 13–15 and Recital 71, and what you find is a requirement to provide “meaningful information about the logic involved” — with genuine, ongoing debate among regulators and legal scholars about what that actually means in practice. GDPR doesn’t even explicitly grant a “right to explanation”; it’s inferred. The framework was contested before modern AI arrived. Large models didn’t break a clean framework; they stress-tested an ambiguous one.

That matters because, in the absence of clear guidelines, the standards organizations actually have to meet are believability and traceability. Can you credibly describe how the system reached its decision? Can you trace the data, the controls, and the human checkpoints? Have you documented it clearly enough to walk a regulator, a customer, or a court through it without flinching? That’s the real test today.

So no, I don’t think we’re enforcing laws that no longer fit. We’re operating in a gap that regulators and industry need to close together. Until they do, the burden is on companies to set their own bar: traceable data, auditable decisions, guardrails on outputs, and documentation you’d be comfortable defending.

AI works best with lots of data, but privacy rules call for using as little data as possible. If companies have to choose, will they prioritize performance over principle? Are we already seeing this happen?

There’s real tension here, but the framing as a binary choice between performance and principle is a bit limiting. The premise that AI works best with “lots of data” is itself worth challenging. More data isn’t automatically better — if it’s poor quality, incomplete, or stripped of the right context, you’re just feeding the model noise. And noise-in produces worse-outcomes-out: hallucinations, bias amplification, and decisions you can’t defend. I’d rather have a smaller, well-governed, well-contextualized data set than a sprawling lake of mixed-quality inputs, while certainly following the GDPR tenet of Privacy by Design.

I think it reframes the privacy question. Privacy rules pushing companies toward data minimization aren’t necessarily working against AI performance — in many cases, they’re forcing the discipline that actually improves it. The companies getting this right are being deliberate about their data strategy: prioritizing quality, relevance, and governance over volume. That’s not a compromise position; that’s just better engineering.

Are we seeing companies cut corners on privacy for short-term performance? Yes, and it tends to come back to bite them through regulatory exposure, customer trust erosion, or models that don’t generalize the way they thought. Trust is becoming a genuine differentiator, particularly in customer-facing and enterprise use cases, and you can’t retrofit it.

The right answer is to design systems where privacy and performance are engineered in from the start, rather than treated as a trade-off you settle later. When done well, they reinforce each other rather than compete.

With decisions like Schrems II and laws like the CCPA, are we heading toward a split internet where data cannot move freely across countries? If so, what will break first: innovation or trust?

What’s interesting about the question is that it frames the split as a US–Europe divergence, when the more consequential fault line is East versus West — between western frameworks debating how to balance rights and commerce, and an eastern framework where the state’s relationship to data is structurally different. That gap isn’t closing through a successor to the Privacy Shield/US-EU Data Privacy Agreement.

So yes, we’re already in a split internet. Between Schrems II, CCPA, the EU AI Act, India’s DPDP, China’s PIPL, and a patchwork of US state laws, any global business is operating across fifteen-plus regulatory environments. My background on both the technology and legal sides of things, coupled with my ability to adjust to both business and customer needs, makes this isn’t hypothetical anymore — it’s the operating environment. We architect for it: data residency, regional processing, model deployment choices that respect where data can and can’t go.

On what breaks first — innovation and trust fail together, even if one precedes the other. If regulation becomes so prescriptive that nothing can cross borders without months of legal review, innovation slows. If companies route around the rules, trust collapses, and regulators tighten further. It’s a doom loop either way.

The companies that come through this well won’t bet on innovation at all costs over trust, or really cumbersome trust over innovation. They’ll invest in both, and accept that regulatory complexity is now part of the engineering/product/support lifecycle, not separate from it.

The post The CIO Who Says Governance Can Actually Speed Up AI appeared first on MartechView.

Every industry has a word it reaches for when it does not know what else to say. In boardrooms today, that word is transformation. It gets invoked at strategy off-sites, written into annual reports, and handed to consultants as a mandate — often before anyone has agreed on what exactly needs to change, or why.

The results are predictable. Brands that once stood for something begin to feel like everything. Customer experiences that were once a competitive advantage become inconsistent and interchangeable. Revenue targets get hit in the short term. Brand equity quietly erodes over the long term.

The problem, more often than not, is not a lack of ambition. It is a lack of an anchor. Transformation without a clear sense of what must be preserved is not strategy — it is change for its own sake.

Giovanna B. Questioni has spent her career in that gap. A transformation expert with experience across luxury fashion, food, furniture, and the digital industries, she has led mergers and acquisitions, crisis interventions, and large-scale omnichannel overhauls across global markets. Her argument is not against change. It is for coherence.

The brands that survive disruption, Questioni argues, are the ones that understand the difference between what is negotiable and what is not. Design, quality, emotional resonance — these are not variables to be optimized in a transformation roadmap. They are the reason the brand exists. Everything else is in service of them.

That distinction shapes how she thinks about omnichannel strategy — an area where many brands have spent heavily and delivered inconsistently. The instinct is to treat every new channel as an opportunity. The discipline is to treat every channel as a responsibility to serve customers without breaking what they already trust.

“Every touchpoint — whether in retail, franchising, wholesale, e-commerce, or social commerce — must deliver a distinct yet cohesive experience. Physical stores thrive on human connection and sensory engagement, while digital platforms excel through simplicity, speed, and personalization.” 

A customer who buys online and returns in-store should not feel the friction of two separate systems. The seams should never show. And the sales assistant at that moment is not a workaround — they are an opportunity for the kind of personalization no digital platform has yet replicated.

What Questioni warns against is the version of transformation that becomes its own end. “Disruption is powerful, but it must always align with customer expectations and ROI,” she says. “Without this balance, transformation risks becoming an academic exercise — one that could dilute the brand’s reputation and desirability.” Bold innovation is not the enemy of brand integrity. Undisciplined innovation is.

The second argument she makes — and the one that challenges the most deeply held assumptions in corporate strategy — is about resilience. The dominant understanding is reactive: how fast can an organization absorb a shock and return to normal?

That framing, Questioni argues, is entirely the wrong one. By the time a company is managing recovery, it has already lost the most valuable thing: time. The organizations that emerge from disruption stronger are not the ones that responded fastest. They are the ones that had already built for it.

“Resilience isn’t just about bouncing back — it’s about building the future before it arrives. It’s the difference between surviving the unexpected and shaping it into opportunity.” 

A future-ready organization, in her view, has designed agility into its operating model before crisis arrives — and has given its teams the tools, authority, and mindset to move without waiting for direction from above. The question she puts to every leadership team is not how they responded last time. It is what they have already built for next time.

Which brings the conversation to the question that defeats most large-scale transformations before they ever reach the customer: execution. The strategy is rarely the problem. The problem is that a vision designed at the top must be delivered by teams moving at different speeds, across different functions and geographies.

Most organizations respond with governance frameworks. Questioni’s response cuts closer to what the problem actually is.

“Large-scale transformation operates like a symphony — an intricate performance where each element must harmonize under the guidance of a skilled conductor.” 

The conductor she has in mind is not ceremonial. It is a deeply operational C-level leader who holds the granular and the panoramic simultaneously — who understands how individual performance feeds collective outcomes, and who can align diverse functions without erasing what makes each of them effective.

The failure mode she sees most often is not a shortage of talent or resources. It is teams executing brilliantly within their own lanes while the overall composition falls apart. “True transformation isn’t about managing chaos,” she says. “It’s about precision, collaboration, and leadership that turns vision into reality.”

The conductor is not optional. Neither is the score. And the brands that understand the difference between noise and music — between change and transformation — are the ones that will still mean something when the disruption settles.

The post The Organizations That Survive Disruption Never Had to Recover From It appeared first on MartechView.

In February 2024, OpenAI gave a small, selective group of creative professionals early access to Sora, its text-to-video generation tool. The list was short. Nik Kleverov was on it.

That alone tells you something. Kleverov is the Chief Creative Officer and co-founder of Native Foreign, an Emmy-nominated Los Angeles creative agency that has spent the last several years building what it calls AI Labs — a practice dedicated not to experimenting with generative tools, but to embedding them as foundational infrastructure across every stage of production. The agency designed the Narcos title sequence for Netflix. It produced the first brand film ever made using OpenAI’s Sora, for Toys”R”Us. It has won twelve Davey Awards for creative innovation in emerging tech.

When Kleverov talks about AI in advertising, he is not theorizing. He is reporting from work.

And what he is reporting is that most of the industry is asking the wrong question.

The Infrastructure Argument

The dominant conversation about AI in creative agencies centers on efficiency. How much faster can we concept? How much cheaper can we produce? How many rounds of iteration can we compress into a single afternoon? Those are real questions with real answers — but Kleverov argues they miss the point, and that the miss is costly.

“The biggest cost is thinking of AI as a speed hack instead of a creative system,” he says. “When it’s treated like a shortcut, there’s novelty and surface-level savings, but not strategy. When it’s treated as infrastructure, it changes how ideas are developed, prototyped, and executed from day one.”

The distinction is architectural. A shortcut is something you reach for after the idea already exists. Infrastructure is what the idea is built on. The difference between the two isn’t visible in the output of a single campaign — it becomes visible over time, in the compounding gap between agencies that have rebuilt their creative process around AI and those that have not.

For CMOs navigating vendor conversations, Kleverov frames the test simply: “The honest question is: do you have an AI workflow, or are you just tinkering? There’s a big difference between occasionally using generative software and actually rethinking how ideas move from concept to production. The companies that treat it as infrastructure will move faster, think bigger, and leave their competition in the dust.”

The Carl’s Jr. Case Study

The clearest recent illustration of the Native Foreign approach is the Carl’s Jr. campaign featuring Paris Hilton — a piece of work that required Kleverov’s team to make a series of deliberate decisions about where AI entered the process and, just as deliberately, where it did not.

The campaign, created using Freepik, leaned into Hilton’s early-2000s cultural moment, framing her as a boss overseeing an AI-automated version of the famous Starwash. The conceit was precise: in the age of AI, Paris Hilton uses it to run her operation — while she oversees it. The nostalgia was not incidental. It was structural.

“We didn’t just prompt ‘2000s aesthetic’ and call it a day,” Kleverov says. “We studied the textures, lighting, and slightly over-the-top tone of that era’s advertising and rebuilt it intentionally — but for today’s audience. Nostalgia works when it feels like memory.”

That distinction — memory versus algorithm — is where the campaign’s creative judgment lived. AI played a significant role in the production workflow, handling elements that could be generated, iterated, and refined at scale. But the scenes featuring the real, present-day Paris Hilton were kept entirely outside the AI pipeline. The boundary was not arbitrary. It was a considered decision about where human presence creates irreplaceable authenticity and where generative tools could amplify without distorting.

What Narcos Taught Him About AI

Before the AI era, Kleverov’s most celebrated work was the title sequence for Narcos on Netflix — a piece built on traditional motion design mastery, painstaking in its craft, executed without a single generative tool. It remains one of the most recognized title sequences in recent television history.

The skills that made that work possible are, Kleverov argues, more valuable now than they were then — not less. Not because AI hasn’t changed the game, but because AI has changed it in a specific way that makes certain human capabilities more critical rather than redundant.

“Tools can generate infinite options, but knowing what not to use has become the real creative skill,” he says. “The fundamentals of storytelling, pacing, and design judgment still act as the compass. Especially with AI storytelling.”

The infinite options problem is one that anyone who has spent serious time with generative tools will recognize immediately. The bottleneck in AI-assisted creative work is not generation — it is selection. Producing a hundred viable options takes seconds. Knowing which one is worth developing, and why, requires everything that cannot be prompted.

What Creative Directors Must Unlearn

That shift has implications for how creative leadership itself needs to change. Kleverov is direct about what the transition demands from Creative Directors who want to work effectively with AI as a foundational layer.

“Creative Directors have to let go of the idea that the first good idea is the one you execute,” he says. “AI rewards exploration, iteration, and divergence. The job becomes less about protecting a single concept and more about guiding a field of possibilities toward the strongest story.”

This is a significant unlearning. The traditional creative director role was built in part around the conviction and the authority to champion a single idea against the instinct to dilate, hedge, or over-iterate. That conviction remains valuable. But the context has changed. When iteration is cheap and divergence is generative rather than dilutive, the skill set shifts from protection to navigation — from defending the best idea to finding it within a field that AI has made vastly larger.

The Democratization Question — With a Caveat

The argument that AI is democratizing high-end production is one Kleverov partially accepts. More people can now make things that look impressive on the surface. Access to tools that once required significant budgets and specialist teams has broadened meaningfully. That is real.

But it comes with a structural caveat. “The gap between something that looks good and something that’s culturally resonant is still huge,” he says. “If anything, taste matters more than ever.”

This is the counterintuitive consequence of democratization in creative industries: as the floor rises, the ceiling becomes the only differentiator that matters. When every agency can produce visually polished work using the same generative tools, the question is no longer whether you can make something beautiful. It is whether you can make something that means something, and that question has always been answered by the same thing it was answered by before AI existed.

Judgment. Context. A point of view that no model was trained to have.

That is what Native Foreign is selling. And if the Carl’s Jr. campaign, the Toys”R”Us film, and the Narcos sequence are any indication, it is a point of view worth listening to.

The post AI Won’t Save Your Campaign. Your Taste Will. appeared first on MartechView.

Raj De Datta has spent over a decade building Bloomreach into one of the most quietly formidable companies in commerce technology. With $260 million in ARR and customers like American Eagle and SPANX, the platform’s AI engine — Loomi — isn’t a bolt-on feature. It’s the whole architecture. But as generative AI rewrites the rules of how people shop, discover products, and interact with brands, De Datta is less interested in celebrating how far things have come and more focused on what comes next. 

In this conversation, he makes a case for why the traditional e-commerce site is becoming infrastructure rather than interface, why personalization and automation are converging faster than brands are ready for, and why the winners of the next era of commerce won’t be those who automate the most — but those who automate responsibly.

Excerpts from the interview; 

Bloomreach just crossed $260 million in ARR, powered by Loomi AI and brands like American Eagle and SPANX. What did you get right in scaling an AI-native platform — and what nearly broke along the way?

We got a few core decisions right early. First, we built Loomi AI as the foundation, not as a feature. It is the intelligence layer that powers our applications and agents across email, search, personalization, and conversational shopping. That architecture meant that as AI advanced rapidly over the past few years, we remained ahead of the curve.

Second, we focused on real-time data and first-party context. Loomi AI combines customer and product data with real-time infrastructure, AI decisioning, and orchestration across channels. That allowed us to deliver personalization that compounds — every interaction feeds the system and improves the next one.

Third, we invested early in agentic capabilities. Nearly half of our customers now use at least one next-generation AI tool, and that adoption has more than doubled in the past year. We also saw strong engagement with our conversational shopping agent, including a significant uptick during the recent holiday season. That momentum reflects a clear product direction: move from static tools to systems that can take action.

The one ongoing challenge — though really more of an opportunity — was the pace of change in AI. Every day unlocked a new possibility. We constantly had to examine what we were building to ensure it was maximizing AI’s potential. That meant a lot of rapid iteration and required us to be willing to take products we thought were great and challenge ourselves to make them even better.

The past two years have rewritten the rules of AI almost quarterly. How do you build a durable product strategy when the underlying technology shifts this fast?

The technology will keep changing, and our strategy is to stay one layer above it. We built Loomi AI as the intelligence layer beneath all of our applications and agents. Because that foundation is consistent, we can adopt new AI capabilities as they emerge without rebuilding the core each time the model landscape shifts.

Durability comes from that architecture — every interaction feeds back into the system, so the platform continuously improves. The goal isn’t to predict the next breakthrough. It’s to build a system that seamlessly incorporates it. That’s what allows us to move quickly and keep innovating while maintaining stability for enterprise customers.

You’ve spoken about agentic commerce. What changes when AI stops assisting shoppers and starts transacting on their behalf?

When AI moves from assisting shoppers to acting more autonomously, commerce shifts from reactive experiences to outcome-driven ones. Instead of simply answering questions or suggesting products, AI can understand context, predict intent, and guide the entire journey in real time — from discovery to decision to purchase.

But this changes a great deal for brands. It means they risk losing control over customer journeys and relationships. That’s why this is such an inflection point for businesses right now. The brands actively investing in agentic commerce — preparing catalogs for agentic discoverability, building their own apps — are the ones that will be prepared for this new era of commerce.

If AI becomes the primary interface for product discovery, what happens to the traditional e-commerce site? Is the homepage becoming obsolete?

It’s more a case of “the website is dead… long live the website.” In reality, the website becomes infrastructure rather than interface, as we know it today. All of the data it houses — about your products, about your customers — remains as relevant as ever. But that data and infrastructure will transcend the website, too. It’ll extend into conversational experiences in AI apps, and into channels like email and mobile, as it already does today. The homepage isn’t obsolete. It just isn’t the front door anymore.

Every AI transformation has its blind spots. What assumptions about AI in commerce turned out to be wrong?

There were assumptions about speed and scale that were somehow both too big and not big enough. At the onset of generative AI and large language models, there was real fear that e-commerce would become irrelevant almost overnight. Obviously, that hasn’t been the case.

But at the same time — it will, and already is, wholly transforming what commerce looks like. The way people discover products, compare brands, build wardrobes — all of that is now done with AI in ways we couldn’t have imagined a few years ago. And I think the scale of that transformation hasn’t even peaked yet.

At what point does personalization cross into automation? And how do brands ensure agency remains with the consumer?

Personalization becomes automation the moment systems start acting without asking. That’s the real inflection point. For years, personalization meant better recommendations or more relevant messaging. Now, with agentic AI, systems can execute decisions across channels. The question isn’t whether that’s possible — it’s how it should be governed.

Brands need to be clear about intent. Automation should reflect what a customer has already signaled, not replace it. When AI is grounded in real customer context and real-time signals, it can reduce friction and help people complete their journey. When it drifts from that, it stops being helpful and starts becoming opaque.

Agency remains with the consumer when technology is designed to respond to their intent, stay transparent in its decisioning, and keep humans in control of outcomes. The more autonomous systems become, the more important those guardrails are.

In this next phase of AI, the differentiator won’t be who automates the most. It will be who automates responsibly — and keeps the customer at the center of that automation.

In 2030, what will feel outdated about how we shop today — and what will Bloomreach have built to stay ahead of that shift?

Today, much of e-commerce still centers on browsing: navigating pages, filtering options, and manually managing campaigns across channels. But the shift we’re already seeing is toward agentic experiences — systems that understand intent and take action in real time. As AI moves from insight to execution, customers will expect more conversational, personalized, and responsive interactions, not just better search results or product recommendations.

What will matter most is whether brands can operate across touchpoints as one connected system. Experiences won’t be confined to a single website or channel. They’ll need to work across email, messaging, mobile, search, and beyond — with intelligence that adapts instantly to customer context.

The post Your Homepage Isn’t the Front Door Anymore appeared first on MartechView.

For years, AI in retail was framed as an efficiency play: better inventory forecasting, smarter demand planning, faster ticket resolution. It was about reducing costs, shaving seconds off handle times, and streamlining back-end systems.

That era is over.

Today, AI sits much closer to the revenue engine. It influences how customers discover products, how they interact with brands, how issues are resolved, and, increasingly, whether they return. In retail, travel, and hospitality, where loyalty is fragile and competition is relentless, customer experience is no longer a support function; it is a core function. It is a strategy.

Few executives have observed that shift from both the operational and technological sides as closely as Michael Klein, Director of Retail, Travel & Hospitality Product Marketing at Talkdesk. Before moving into enterprise CX technology, Klein spent more than three decades in retail merchandising and leadership roles, including time with Adobe and Williams-Sonoma.

That background, he says, fundamentally shapes how he evaluates enterprise software.

“I’ve spent more than three decades in retail,” Klein said. “My background in retail merchandising keeps me focused on how technology actually improves the customer experience and drives real business outcomes. It’s never been about leveraging technology just for the sake of it.”

In merchandising, every decision — from assortment planning to store layout — is tied to measurable results. Klein brings that same lens to AI.

“AI is a strong example,” he said. “When applied well, it helps increase average order value, drive repeat purchases and improve retention.”

In other words, the question is not whether AI is impressive. It is whether it sells more sweaters, books more rooms, or deepens loyalty.

“At Talkdesk, we’re focused on using it where it truly makes a difference for retailers and their customers,” he added.

The Persistent Myth of the Demographic Customer

Retailers have more customer data than ever before. Yet, paradoxically, many still design experiences around broad assumptions.

“One of the biggest misconceptions is that everyone in a demographic bucket behaves the same way,” Klein said. “That Gen X shops one way, Millennials another, and Baby Boomers another.”

The reality, he argues, is messier.

“If it were that clean, personalization would be easy,” he said. “But I’ve seen plenty of Baby Boomers who are perfectly comfortable booking travel online and plenty of younger customers who want to speak to a person when something goes wrong.”

The mistake, he suggests, is building customer journeys around stereotypes rather than behavior. Digital-first consumers are not defined by age alone; they are defined by context, urgency, and preference at any given moment.

Designing around assumptions creates friction. Designing around actual signals creates loyalty.

From Cost Center to Growth Engine

Perhaps the most profound shift underway is the redefinition of the contact center.

For decades, contact centers were treated as overhead — necessary to resolve complaints, but disconnected from growth. AI is changing that calculus.

“From a customer standpoint, AI makes it easier to self-service when they want to,” Klein said. “That convenience drives satisfaction — and satisfied customers tend to spend more.”

Self-service is not about removing human interaction; it is about giving customers control. When done well, automation reduces frustration and accelerates resolution.

For brands, the internal impact is just as significant.

“AI frees up contact center teams to focus on higher-value work,” Klein said. “Whether that’s helping design a room, building a wardrobe, or solving a complex issue. Those are real revenue-driving conversations.”

By removing friction in knowledge access, documentation, and agent training, AI shifts human effort toward consultative interactions — the kinds of conversations that build trust and increase basket size.

At its best, automation does not replace people. It elevates them.

Why So Much “AI-Powered CX” Falls Flat

In a market saturated with “AI-powered” claims, differentiation often dissolves into jargon. Klein has spent years translating complex enterprise technology into language retailers actually understand — and he is blunt about what works.

“Great marketing and complex jargon just don’t mix,” he said. “Instead of relying on technical words that only developers or product managers use, brands should drop the IT acronyms and language that’s too deep in the weeds.”

Retailers do not buy platforms. They buy outcomes.

Plain language, he argues, forces vendors to clarify their value proposition. If a product cannot be explained without abstraction, it likely lacks practical grounding.

Separating Real AI From Marketing Noise

Klein is equally pragmatic about the current AI moment. Not all AI is new — and not all of it is transformative.

“We first need to be clear about which AI we’re talking about,” he said.

Retailers have long used predictive models to forecast inventory, optimize distribution, and manage replenishment. Those systems quietly shape customer satisfaction by ensuring products are in stock.

“That has a direct impact on product availability and customer satisfaction,” Klein noted.

More recently, generative and agentic AI have begun to deliver tangible value in marketing and service environments — drafting content, assisting agent,s and streamlining workflows.

Where the narrative drifts into hype, he says, is in the idea of full automation.

“We’re a long way from AI taking over everything,” Klein said. “Human oversight still matters, and consumers will want the choice between automation and a real person depending on the situation.”

The future, in his view, is hybrid — not robotic.

The Context Problem

If there is one area where brands consistently misstep, it is context.

With so much data available, companies often mistake volume for insight.

“The key to designing better experiences is relying on context signals, like timing and intent, coupled with history and preference,” Klein said.

Without context, personalization becomes misdirection. A customer who once purchased a gift for a relative may be permanently misclassified, leading to irrelevant recommendations.

“Imagine you visit a store for the first time to buy a present for your grandmother,” he said. “If the retailer caters your experience based only on your first visit, your experience won’t serve your current needs.”

The difference between intelligent personalization and awkward irrelevance often comes down to whether brands understand why a purchase occurred — not just that it did.

The Hardest Part of Modernization

Enterprise modernization is rarely blocked by technology alone.

“The biggest hurdle brands encounter is not doing anything for fear of disrupting operations,” Klein said.

Legacy systems may be aging, but they are stable. Change introduces uncertainty.

“The next hurdle is weeding out bad data,” he added. “And the third is dealing with people who are stuck in their old ways or too protective of their territory.”

Transformation demands both technical cleanup and cultural shift. It requires encouraging teams to experiment, test openly, and share ownership across departments.

Without that alignment, even the best technology stalls.

The Metrics That Will Matter Next

As AI becomes embedded in CX systems, traditional metrics such as handle time and surface-level satisfaction scores may lose their primacy.

“In the next few years, customer lifetime value, recency, and frequency will become key metrics to monitor,” Klein said, particularly in retail and hospitality.

Those measures capture relationship strength rather than transaction speed. They reflect whether AI is driving durable loyalty rather than short-term efficiency.

In that sense, the next phase of AI in customer experience is not about the volume of automation or technological sophistication. It is about commercial impact.

For Klein, that is the through line connecting his merchandising past to his product marketing present. Technology is not the hero of the story. The customer is.

And in a competitive landscape where switching costs are low and expectations are high, the brands that treat customer experience as a growth lever—not a service line item—will be the ones that endure.

The post Retail’s AI Reckoning Is About Revenue — Not Robots appeared first on MartechView.

In Silicon Valley, obituaries are often written too soon. First, it was email. Then the press release. Now, the target is the public relations agency itself.

The argument is seductive in its simplicity: if artificial intelligence can draft a pitch, summarize a strategy, and produce passable thought leadership in seconds, what need remains for the humans who were once billed by the hour to do it?

To test that thesis, I sat down with Susan Thomas, CEO and founder of 10Fold, a woman who has guided more than 500 companies toward the grueling scrutiny of the public markets. Her answer was neither defensive nor nostalgic.

“Traditional PR is dying,” she told me. “That does not mean media outreach or thought leadership campaigns are going away any time soon. What I mean is the traditional boundaries and limitations of PR no longer apply and firms must adjust quickly to adapt.”

It was not a lament. It was a warning.

For decades, public relations operated under a comfortable ambiguity. Visibility was presumed valuable. Influence was inferred. Impressions were counted and filed away in quarterly reports. In a world before dashboards and CRM pipelines, that was enough.

It is no longer enough.

The rise of measurable marketing—SEO, performance advertising, revenue analytics—has forced communications into a harsher light. Chief executives now ask questions that clip books cannot answer: How much pipeline? How much attributable growth? What business outcome changed because of this campaign?

Thomas believes the industry’s real reckoning is not with AI, but with accountability.

“AI is not the existential threat to PR and communications agencies,” she said. “To survive in today’s volatile environment, they simply have to earn their place. That means having client discussions that begin with business objectives, which serve as the basis for a communications plan. It’s about reverse engineering the desired business outcomes.”

A press release, she added bluntly, “has no inherent value.” It is a cost unless it drives measurable movement — website traffic spikes, CRM engagement, investor attention. At 10Fold, she said, teams correlate coverage with direct traffic patterns, AI rankings and campaign timelines. It is not perfect, she admits, but it is far more rigorous than impressions and advertising equivalents, the old currency of the trade.

In this framing, AI is not the villain. Mediocrity is.

The AI Paradox: Speed vs. Substance

The fear animating many agencies is that large language models have commoditized content. Thomas sees it differently.

“AI adoption is not killing agencies; it is critical for both agencies and companies alike,” she said. “The problem is that most don’t use it correctly. AI relies on data and input — something very hard to find when you have an original idea, solution, or approach.”

AI, she argues, is a powerful thought partner and time saver. But it draws from what already exists. Without original insight and a defined position, it produces content that is technically fluent and strategically hollow.

“Without a strong human point of view,” she warned, “AI-generated content quickly becomes generic and indistinguishable.”

In a discovery ecosystem increasingly shaped by large language models that prioritize credible third-party validation, indistinguishability is invisible.

The IPO Illusion

Thomas has seen another recurring misstep: high-growth technology companies failing to evolve their narrative as they move toward an IPO.

“The biggest mistake executives make,” she said, “is not understanding how communications plans must evolve as the company matures and moves toward an exit.”

Early-stage companies talk about innovation and disruption. But as they approach public markets, the narrative must broaden. Investors look for ecosystem relevance, customer validation, and operational discipline.

Hiring a CFO can signal financial maturity. Strategic partnerships signal integration. Vertical expertise signals durability. And once financial institutions are secured and paperwork is filed, the quiet period reshapes what can and cannot be said. Even subtle language shifts in press materials become regulated terrain.

Companies that fail to prepare in stages, she noted, often find themselves constrained at precisely the moment clarity matters most.

The Measurement Reckoning

Public relations has struggled to quantify its impact, not because it lacks impact, Thomas argues, but because it failed to adopt modern measurement frameworks.

“For much of its 100-year history, PR operated on a simple premise: visibility and influence were assumed to be valuable,” she said. “Anecdotal success stories are not enough in today’s environment.”

Compounding the issue is timing. Reporting delivered months after a campaign’s peak offers little strategic leverage. Measurement, she insists, must be continuous and integrated with the systems marketers trust — website analytics, CRM dashboards, pipeline models.

Agencies that sell “activity as value” will not survive. Those that demonstrate proof points aligned to business goals will.

Investors, Narrative, and the New Discovery Engine

Investors, Thomas said, are not interested in marketing spin. But they care deeply about credible narratives that resonate with buyers, partners, and future backers.

With large language models reshaping how information surfaces, third-party validation has become even more important. Discovery is shifting away from keyword density and toward authority and originality. In that environment, earned media and differentiated thought leadership are not ornamental — they are strategic assets.

Strong brand programs, supported by credible external voices, increasingly serve as signals of long-term value creation.

Crisis in the Always-On Era

Corporate reputation no longer moves in neat cycles. Social media and activist stakeholders compress timelines and amplify scrutiny. Yet Thomas resists the notion of a permanent state of emergency.

Preparation, she argues, is the antidote. Establish baseline sentiment before a crisis. Benchmark normal conditions. Define response protocols in advance.

“When leaders understand where their organization stands in normal conditions,” she said, “they are better equipped to respond with clarity and confidence.”

Beyond the Hype

Having spent decades in Silicon Valley, Thomas has seen hype cycles crest and collapse. She does not place AI among them.

“AI is beyond a hype cycle,” she said. “It is fundamentally transforming business processes at every level.”

The exuberance will normalize. Integration will deepen. The tools will become infrastructure. What will remain scarce is not automation, but originality.

The obituary for traditional PR may well be accurate. But obsolescence is not inevitable. Reinvention is.

In the age of automation, the agencies that endure will not be those that defend legacy practices. They will be those that prove — with discipline, data, and differentiated thought — that narrative, when aligned to business outcomes, remains one of the most powerful assets a company can possess.

Prove your value, or be automated out of the conversation.

The post AI Isn’t Killing PR. Bad Measurement Is. appeared first on MartechView.

In healthcare marketing, ambition is always tempered by responsibility. The industry sits at the intersection of innovation and regulation, where data promises precision but privacy demands restraint. Julius Ramirez, EVP and GM of Global Data & AI Products and Partnerships at Doceree, has built his career navigating that tension—across startups, large-scale tech platforms, and now one of the fastest-growing AI-powered healthcare marketing companies.

At a moment when artificial intelligence is redefining targeting, measurement, and engagement, Ramirez argues that the real transformation is not about speed or scale. It is about architecture. It is about trust. And it is about building systems designed for compliance from the ground up, rather than retrofitting ethics after the fact.

The Maturation of Personalized Marketing

For years, personalized marketing relied on what Ramirez calls “convenient data”—signals that were easy to access, loosely governed, and often detached from real consumer intent. That era, he suggests, is ending.

Privacy regulations are tightening globally, and healthcare has always operated under stricter scrutiny than most industries. Rather than seeing this as a constraint, Ramirez views it as a forcing function.

“Personalized marketing is maturing,” he explains. “The bar is higher. You need durable identity frameworks, consent-led data strategies, and partnerships built for compliance from day one.”

To some, this sounds expensive. To Ramirez, it is a reallocation of value. Smarter identity resolution and governed data flows eliminate waste—duplicated reach, low-fidelity targeting, and signals that should never have driven decisions in the first place. What replaces them is precision and accountability.

In healthcare, where trust directly influences engagement, that precision carries commercial weight.

Beyond Algorithms: An Operating System Approach

Doceree describes itself as the only AI-powered operating system for healthcare marketing—a claim that invites scrutiny. Ramirez welcomes it.

“The difference isn’t in the algorithm,” he says. “It’s in the architecture.”

Many AI platforms in life sciences function as point solutions—optimizing media buying, segmentation, or analytics in isolation. Doceree, he argues, built its system around healthcare’s structural realities: HCP identity, clinical context, consent requirements, and regulated data environments.

Rather than layering AI onto legacy stacks, Doceree embeds intelligence into its decision layer—connecting identity, activation, and measurement in a unified environment.

Healthcare, Ramirez notes, is not simply another vertical. Signals matter differently. Understanding care settings, treatment moments, and professional intent requires contextual intelligence—not just cookies or clicks.

When Doceree calls its platform an operating system, Ramirez says, it means AI governs how data flows and how engagement happens across endemic healthcare environments. The system orchestrates the ecosystem itself.

Ethics as Architecture, Not Afterthought

Few sectors feel the commercial pull of hyper-targeting more acutely than healthcare. Yet few sectors operate under tighter ethical and compliance guardrails.

Ramirez rejects the framing of this tension as a trade-off.

“In healthcare, the guardrails are design inputs,” he says. “They’re not obstacles.”

Doceree’s systems assume sensitivity from inception—de-identified data structures, consent-led frameworks, and strict separation between patient-level signals and healthcare professional engagement strategies. Hyper-targeting, in his view, should rely on context and professional intent—not on exploiting sensitive personal data.

The misuse of data in healthcare does more than invite regulatory risk. It damages credibility with providers and partners. That reputational cost, Ramirez argues, far outweighs any short-term performance boost.

The discipline imposed by compliance, he believes, forces better AI models and stronger long-term outcomes.

Lessons From Big Tech

Ramirez’s time in large-scale technology environments, such as Meta, shaped his thinking about measurement discipline.

In Big Tech, he says, metrics are stress-tested relentlessly. Definitions evolve. Attribution models are debated. Nothing is taken at face value.

By contrast, earlier-stage HealthTech and AdTech companies often accept “directionally right” measurement. That tolerance may work in the short term, but ambiguity compounds as companies scale.

“Eventually,” Ramirez says, “scale demands precision.”

The earlier organizations adopt rigorous definitions, feedback loops, and measurable outcomes, the more durable their growth becomes.

Intelligence Versus Automation

AI vendors routinely promise “measurably better outcomes.” Ramirez draws a clear distinction between automation and true intelligence.

Automation executes predefined logic more efficiently. AI, by contrast, should refine its own logic.

“If a system just adjusts bids or frequency caps faster, that’s efficiency,” he says. “Intelligence means it learns.”

True AI performance manifests in adaptive decision-making—systems that ingest new signals, continuously update models, and stabilize performance even as environments shift. The proof lies not in surface metrics such as click-through rates but in downstream behavioral changes and predictive accuracy over time.

Intelligence, in Ramirez’s definition, is adaptive and accountable.

The Partnership Question

In the crowded data and AI landscape, partnerships are ubiquitous. Many, Ramirez acknowledges candidly, amount to little more than logo slides.

“A real partnership changes capability,” he says. “It alters how data flows, how identity is resolved, how activation or measurement works.”

Operational partnerships require shared standards, technical integration, governance agreements, and often joint roadmaps. They are slower and more difficult to execute. But they create compound advantage.

Logo swaps, he suggests, depreciate as quickly as the press release cycle that announces them.

Five Years Ahead

Looking forward, Ramirez does not see regulation slowing healthcare AI. Instead, he anticipates transformation—but not the theatrical kind.

The shift will be toward what he calls “decisioning intelligence”: systems that understand clinical context deeply enough to determine when and why engagement drives action. Healthcare marketing, he predicts, will move away from volume-driven outreach toward precision infrastructure.

Regulation will shape the path but not halt progress. The winners, he argues, will be companies that embed compliance, privacy, and explainability into their architecture from the start.

AI in healthcare marketing will not resemble disruption theater. It will look like disciplined engineering.

And in an industry built on trust, that may be the most transformative shift of all.

The post Healthcare Marketing’s End of “Convenient Data” appeared first on MartechView.

As automotive AI grows more expressive, the challenge is no longer whether in-cabin assistants can feel human—but how far they should go. In this conversation, Olga Khryapchenkova, Lead Product Manager for AI at NIO, explores the delicate balance between emotional connection and safety in a high-stakes environment. From designing culturally adaptive voice interactions to ensuring new AI features are discoverable, Khryapchenkova offers a clear-eyed view of what responsible, user-centered AI looks like inside the vehicle. Her perspective is grounded, pragmatic, and notably free of hype—focused less on futuristic promises and more on building AI that drivers can trust.

Excerpts from the interview; 

As NIO gives its AI a “face,” how do you balance emotional connection with the risk of false intimacy in a safety-critical environment?

A visual presence creates a natural connection to both the product and the brand, which helps with perception and adoption. More importantly, it clarifies system states—when the assistant is listening, speaking, or processing—which directly supports efficiency and driver safety.

The balance lies in designing a warm, purposeful persona that enhances guidance and clarity, without drifting into emotional cues that could create unintended expectations or dependencies.

With AI now capable of tone, humor, and personality, how do you ensure cultural authenticity across markets rather than a one-size-fits-all persona?

It always starts with user research. You have to meet users where they are—listen to their pain points, collect feedback across channels, and combine those insights with strong market understanding before entering a new region. Cultural authenticity isn’t something you retrofit later; it’s something you validate early and continuously.

How do you build an emotionally aware voice assistant while meeting strict automotive demands around latency and safety?

True emotion detection isn’t on the table yet. What is feasible is strong context handling—smooth, multi-turn conversations within defined domains. That keeps interactions natural and even enjoyable while remaining efficient and compliant.

Latency is still a market-wide challenge, though progress is steady. And transparency is essential: users need clear controls and a clear understanding of what’s being captured and why.

GenAI doesn’t localize well out of the box. How do you approach multilingual, culturally adaptive voice interactions at scale?

Again, user research is central. You validate cultural nuances with real users and native speakers, remain agile, and are ready to adjust post-launch when issues surface. Over-the-air updates are invaluable here.

The principle is simple: open-minded, user-centered iteration at scale.

How do you ensure users discover and understand new AI features inside the vehicle?

Features that aren’t surfaced effectively quickly become dead features. Clear release notes are essential, and sometimes Q&A sessions or short explanatory videos help.

Equally important is collaboration. Product managers need to work closely not just with engineering, but with product marketing, go-to-market, and communications teams. Tracking adoption and engagement tells you whether the message is landing—and where it needs refinement.

How do you use user feedback to improve in-cabin AI features?

Feedback comes from multiple channels: the voice assistant itself, the companion app—within a closed-loop system—and surveys. Continuous feedback highlights friction points, while analytics reveal which features need better visibility.

That combination allows product and marketing teams to iterate quickly and communicate improvements in a way users actually notice.

Looking ahead to 2026, what excites you most about AI products?

Smarter base models, more refined AI user experiences, and richer multimodal interactions. That said, we’re still far from true productivity breakthroughs, and AGI remains a distant horizon.

There’s a lot of work ahead—which means we certainly won’t be bored in 2026.

The post Designing AI That Feels Human—Without Crossing the Line appeared first on MartechView.

In the rapidly shifting landscape of customer experience, Adrian Swinscoe remains a steadfast voice for clarity over clamor. As an independent advisor and the author of Punk CX, Swinscoe has spent years advocating for a return to business basics in an industry often blinded by the glitter of new technology.

As we enter 2026, the marketing world finds itself at a crossroads: balancing the breakneck speed of “Agentic AI” with the timeless need for genuine human connection. I sat down with Swinscoe to discuss his predictions for the year, the “Michelangelo” approach to technology, and why he believes the term “CX Expert” is a fundamental contradiction.

As we look at the marketing and martech landscape of 2026, what three shifts do you predict will define the year?

First, we are facing a massive noise problem. The marketplace is saturated with vendors making increasingly grand promises. For brands, the challenge is no longer just communicating; it’s differentiating. There is a lot being said, but very little signal.

Second is the rise of Answer Engine Optimization (AEO). Whether it’s through Gemini, ChatGPT, or Perplexity, the way customers find information has fundamentally shifted. Brands have to figure out how to gain visibility within these AI-driven interfaces and how that integrates with their broader marketing strategy.

Finally, there’s the economic reality. Conditions aren’t getting easier for the average customer. In a tighter economy, generating attention and loyalty is harder than ever. Brands must return to the basics: doing what you say you’re going to do, when you say you’re going to do it. Consistency is the new premium.

Every year has its buzzword. What is the most overhyped term in CX right now, and what is the reality behind it?

Without a doubt, it’s “Agent AI.” While it offers massive potential, we have to remember it’s only a tool. Most brands have significant “housekeeping” to do—data hygiene, infrastructure, governance—before they can actually use it.

Think of Michelangelo’s David. It’s a 17-foot masterpiece carved from a single block of marble. Michelangelo had tools, but the tools didn’t make him a master. It was the combination of his vision and his mastery of the medium. Brands need to stop chasing the tool and start mastering their vision and their data.

Many brands rush to layer on AI without fixing their core issues. What is the one element every CX stack must get right first?

You have to prove its worth and its relevance. It sounds simple, but CX leaders have struggled with ROI for years. If your technology doesn’t tie directly to the core levers of the business—generating demand, converting customers, or improving operational efficiency—it’s just a “nice to have.”

I use a simple model: Business is about demand, conversion, and retention. Most people focus on the outcomes—revenue and profit. But those are just functions of how well you manage the process. If your tech stack doesn’t improve those specific levers, you’re just adding complexity, not value.

We are swimming in feedback—surveys, forms, reviews—yet insights seem rare. How do brands actually close the loop?

The traditional “loop” is broken because survey data is increasingly partial. Response rates are falling, and you usually only hear from the delighted or the enraged. You miss the “silent middle.”

To fix this, brands need to listen more holistically across all channels. It’s about mining real-time data from every interaction and plumbing that insight back into the business immediately. We need to move away from “surveying” and toward “listening.”

You’ve long championed “Punk CX.” What does that rebellion look like in 2026’s hyper-automated world?

Two words: Human Connection. The rebellion is focusing on the relationships between the brand, the customers, and the employees.

Take the contact center. Research shows that while 75% of customers want to self-serve, only a fraction are actually successful. We’re building chat-bots without consulting the “conversation experts”—the agents who talk to customers every day. I call this “Agent Intelligence” vs. “Artificial Intelligence.” When you combine the biological supercomputers (your people) with the tech, you win.

“CX” is now a boardroom staple. What’s your one rule for leaders who want to make it real rather than just a buzzword?

Stop talking about “CX” as a standalone thing. Just stop. Start talking about outcomes. How are you improving life for the customer? How are you making the employee’s job better? How is the business growing? If you focus on those three outcomes, traction follows.

For young professionals entering this AI-driven era, what is your advice for staying relevant?

First, reject the title of “expert.” You cannot be an expert in someone else’s experience. You can only be an expert in creating the conditions for a good experience.

Stay curious. Embrace humility. In the true punk style of Henry Rollins: Question everything. Don’t get sucked into the hype. Read widely—not just business books—and stay focused on the human at the other end of the wire.

On the topic of reading widely, what’s on your nightstand right now?

I’m currently looking at The Art of Noticing by Rob Walker. It’s a fantastic book with 131 exercises designed to help you see things differently. For example, it suggests going on a “photo walk” without a camera. You start “seeing” the shots without the distraction of the device. It builds your curiosity muscle, which is the most important skill to have in an AI world.

The post CX Is Noisy. Human Connection Cuts Through appeared first on MartechView.

Artificial intelligence is everywhere in customer experience—or so the marketing claims. Yet beneath the buzzwords and glossy demos, many enterprises are discovering that “AI-powered transformation” often looks suspiciously like old workflows with new labels.

In this conversation, Kevin McNulty, Senior Director of Product Marketing at Talkdesk, cuts through the hype to examine what’s actually changing—and what isn’t. From the limits of single-agent AI to the overlooked importance of data architecture, McNulty offers a grounded view of how AI is reshaping customer service, marketing, and trust itself—and why the next phase of CX will be defined less by automation and more by orchestration.

Excerpts from the interview; 

What’s the most overhyped belief about AI-powered customer experience that enterprises need to abandon?

The most overhyped belief is that a single AI bot can solve complex customer experience challenges simply by being added to an existing workflow. Many organizations assume that plugging in an AI model will somehow create transformational change. In reality, that approach delivers only marginal gains.

True transformation happens when companies move away from isolated AI models performing narrow tasks and toward multi-agent orchestration frameworks. In those environments, AI systems can reason together, coordinate actions, and execute end-to-end outcomes. If your operating model still depends on humans stitching together fragmented workflows, you’re not transforming customer experience—you’re merely augmenting it.

At Talkdesk, we think about AI not as a feature but as a workforce. And like any workforce, AI requires structure, orchestration, governance, and shared context. Once organizations start thinking this way, they also begin to recognize the need for quality management—setting standards, measuring performance, and ensuring accountability. That shift in mindset is where real transformation begins.

Many brands claim they’re moving CX from a cost center to a growth driver, yet their operating models still resemble 2015. What’s the structural barrier holding the industry back?

It comes down to data fragmentation.

Enterprises have invested billions in digital transformation, but customer data remains scattered across CRMs, ticketing platforms, policy systems, and even individual agent desktops. You simply cannot drive a growth-oriented customer experience if your AI lacks a unified, trustworthy understanding of the customer.

There’s also an uncomfortable truth the industry doesn’t like to acknowledge: AI cannot fix bad data. In fact, generative AI that retrieves the wrong information with great confidence is worse than no AI at all. The solution isn’t to keep layering intelligence on top of broken systems—it’s to fix the underlying information architecture first. Only then can organizations unlock meaningful value from AI.

As AI systems take on more of the customer journey, how should brands design experiences for a world where AI intermediaries increasingly act on behalf of customers?

The first requirement is a common language. If AI agents are going to interact with other AI agents, they need shared protocols and governed, traceable knowledge. We’re starting to see early standards—things like agent-to-agent frameworks—emerge, but we’re still at the beginning.

From a design standpoint, this requires a fundamental shift. When AI agents act on behalf of customers, you’re no longer designing linear journeys. Agents can move faster, multitask, and adapt dynamically, which means experience design becomes less about scripting flows and more about defining decision frameworks.

Humans follow scripts. AI agents follow policies, context, rules, and constraints. That’s what designers need to focus on building.

Another critical concept is outcome unification. Behind the scenes, AI-to-AI interactions will become incredibly complex, involving multiple systems and agents collaborating simultaneously. But from the customer’s perspective, the experience must feel seamless. Even if five agents contributed to an answer, the customer should receive a single resolution, a single narrative, and a coherent outcome—without contradiction.

You’ve argued that humans won’t be replaced anytime soon. Where should AI take the lead in customer service to force the industry to evolve?

Knowledge management is the most logical place for AI to lead.

In customer service, the biggest bottleneck is rarely agent capability or staffing levels. It’s that organizations don’t truly understand or manage what they already know. Knowledge is scattered across call transcripts, internal documents, policy files, and historical interactions—and it’s often outdated the moment it’s documented.

AI can transform knowledge into a living, governed system. It can continuously structure unstructured data, surface gaps, keep content fresh, and proactively suggest improvements.

Take the airline industry as an example. A rare policy question may already have been answered once—buried inside a 15-minute call transcript. AI can surface that answer, formalize it, and ensure the next customer receives an instant response instead of repeating a long, frustrating interaction. That’s where AI can dramatically improve both efficiency and experience.

Industries like insurance are rapidly deploying AI agents at scale. When does efficiency begin to erode trust—and how should companies manage that risk?

Trust erodes the moment AI becomes opaque or unaccountable.

There’s enormous pressure to automate, and that pressure can push organizations to deploy AI beyond what their data quality, governance, or policies can support. Our advice is straightforward: autonomy must match data quality. If your data is incomplete or unreliable, your AI should not be fully autonomous.

Explainability must also be non-negotiable, especially in regulated industries like healthcare and financial services. Customers need to understand when they’re interacting with AI and how decisions are being made. Transparency is essential to maintaining trust.

Human-in-the-loop systems are not a crutch—they’re a safety net. The best AI systems escalate intelligently, not because they’re confused, but because they recognize when confidence drops. Responsible automation has to be designed from the beginning. It cannot be retrofitted after something goes wrong.

What three shifts will define marketing and martech in 2026—and why should brands pay attention?

The first shift will be the rise of agentic website experiences, which will begin to replace traditional SEO as the primary discovery model. Websites will no longer rely on navigation or keyword-driven search as their main entry point. Instead, AI agents will guide visitors through conversations. That fundamentally changes digital marketing—fewer page views, more interactions, and a shift from optimizing for keywords to structuring knowledge for machines.

The second shift is the convergence of the marketing stack and the contact center stack. Historically, marketing has owned the top of the funnel while contact centers managed post-purchase interactions. AI collapses that divide. The same data, knowledge, and workflows can power both, enabling far more automation and continuity across the customer journey.

The third shift is outcome-based pricing. As AI adoption accelerates, organizations want clear proof of value. Pricing models will move away from seats and licenses toward outcomes delivered—tasks completed, issues resolved, value created. That will fundamentally change how ROI is defined and how technology investments are justified.

Every year brings a buzzword. What’s the most overhyped promise in marketing right now—and what’s the reality behind it?

The most overhyped promise is the idea of an all-in-one AI agent that can manage the entire customer journey with a single prompt.

The customer journey is inherently complex. It involves identity verification, intake, troubleshooting, compliance, updates, and personalization—often across multiple channels. A single AI agent cannot reason across all those domains with the level of accuracy and specialization customers expect. It will either oversimplify, hallucinate, or fail when it encounters nuance.

The reality is that effective customer experience requires a team of specialized AI agents, each focused on a specific capability, coordinated by an orchestrator that manages conflicts and unifies outcomes. One agent solving everything is a compelling story—but it’s not a realistic one.

For marketers entering an AI-driven era, what’s one piece of advice to stay relevant in 2026 and beyond?

Break the rules.

We’re at the very beginning of this AI era, and this is the moment to rethink everything. Don’t limit yourself to how marketing worked in the past. Younger marketers, in particular, aren’t burdened by legacy assumptions—and that’s a strength.

Imagine the experiences you’d want as a customer. We’ve all used tools like ChatGPT or Claude. Think about what feels intuitive, helpful, and human—and then think about how to build that. You no longer need deep coding expertise to experiment and create.

Marketing success will increasingly be measured by outcomes, not clicks or forms. Challenge assumptions. Break existing workflows. Invent new ones. This moment feels as transformative as the early days of the internet, and it’s moving just as fast.

There’s a natural instinct to fear AI, just as there was fear when the internet emerged. But history shows that these shifts don’t eliminate opportunity—they expand it. The marketers and brands who lean in early will help define what the future looks like.

The post Human-in-the-Loop Isn’t a Crutch. It’s the Safety Net. appeared first on MartechView.