AI has unlocked real-time personalization, predictive insights and cross-channel orchestration. Yet in 2024, customer experience scores declined for 25% of U.S. companies, while only 7% improved.
The paradox: more tools, worse outcomes. The problem isn’t technology—it’s go-to-market (GTM) models that haven’t kept pace. Overcoming the “attention apocalypse,” marked by shrinking focus and constant distraction, will require leaders to pair AI with human insight and rebuild trust by designing frictionless experiences across every role and channel.
In this discussion, ZS’s Omer Hancer is joined by fellow ZS experts Jaideep Bajaj, Arun Shastri and Maria Whitman to explore where leaders can use AI to rebuild customer experience, trust and loyalty.
Omer Hancer: LLMs are evolving fast. Claude Opus 4.1 and GPT-5 have pushed technical boundaries, yet customer experience (CX) scores keep slipping. Why aren’t these capabilities translating into better outcomes?
Arun Shastri: Fundamentals are being neglected. Last year, 25% of U.S. companies saw CX decline because they lost sight of the basics: solving real pain points, mapping the full journey and turning complaints into loyalty moments. Every time there’s a new technology, we think it’s going to solve the way we engage with the customer, and somewhere along the way we abandon the core principles of what makes for good customer engagement.
Recently, a large telecommunication company’s CX transformation only took off when the CEO made net promoter score (NPS) a companywide goal. That forced marketing, sales, product and operations to work together to remove friction across the journey, shifting from isolated fixes to “happy customer journeys,” end-to-end.
Maria Whitman: AI can make experiences faster and more accurate, but the real leap will be in empathy and inclusiveness—not just “human in the loop,” but more human. The fastest growth often sits in underserved segments: elderly patients who struggle with visual interfaces, or retail buyers making high-stakes purchases where anxiety, not price, is the barrier. Imagine AI that switches from visual to voice for an older user or adapts its tone when it senses stress in a transaction. Meeting people where they can best engage opens markets competitors can’t reach and builds loyalty they can’t replicate.
OH: Reducing friction has always been a CX goal. In a modern GTM model, can AI move us from fixing problems faster to preventing them altogether?
MW: Absolutely—and prevention is where AI creates a real competitive advantage. In pharma, it could mean anticipating a doctor’s likely question and proactively delivering the answer before the meeting. In retail, look at how Warby Parker’s virtual try-on removes uncertainty from buying glasses, which is a high-intent, identity-defining purchase. These moments aren’t just immersive; they shorten sales cycles, boost conversions and prevent drop-offs.
Jaideep Bajaj: Prevention only works if you have deep context, not just surface-level preferences. In healthcare, that’s knowing a clinic’s workflow bottlenecks or anticipating a policy change that could disrupt prescribing. Success hinges on meeting customers’ needs without friction. Imagine a rep walking in with a “digital twin” of their medical science liaison (MSL) colleague and being able to have the digital twin answer any question instantly even if the MSL isn’t there. That’s orchestrating the right expertise at the right moment, so the customer keeps moving without friction.
OH: You’ve talked about an “attention apocalypse” with attention spans shrinking across industries. What does that mean for GTM?
AS: For years, we’ve built marketing strategies assuming customers would stay engaged long enough to make deliberate decisions. That’s no longer true. Now, customers—especially younger ones—abandon interactions midjourney. The solution isn’t to push harder. It’s to design short, high-value moments and even allow for “graceful exits” so customers want to re-engage later. That could mean creating short microbursts of activity, then giving them segmented offers a little at a time.
Some experts are predicting that you might need to enable customers to abandon engagement with you on their terms because, if you keep them in that engaged mode and they don’t like it, they won’t come back. This may also mean that companies focus on forming habits rather than driving decisions. So, I think we are going to see remarkably different approaches emanate from the way the next generation engages with media.
MW: When we look at things like online videos and digital advertising, Generation Z gives you about eight seconds to prove relevance; boomers, around 20. For HCPs, who tend to consume digital content in short bursts, the average attention span for a video is less than two minutes. And among patients, we see four times the engagement for story-driven videos compared to brochures.
People expect the same relevance, speed and personalization from pharma that they experience from Netflix, Amazon or Spotify. Meeting their expectations means shifting from maximizing touch points to maximizing impact per interaction. It also means showing up where trust already exists.
In pharma, peer influencer programs have reached patients traditional GTM missed. And sometimes disengagement is intentional. My Generation Z daughter takes multiday breaks from social media to avoid overload. Leaders need to design GTM models that fit those rhythms, whether in consumer or B2B markets.
There are also a lot of things that we must do differently to engage and figure out where people are placing their trust. I’d love to say the answer is responsible AI everywhere, but we know that’s not going to be true. And we also know that the more we use AI, the more people’s content will filter them toward specific things. Fifty percent of Generation Z consumers trust peer-generated content more than any brand-generated content. So, marketers are going to have to step back and say, “Where is trust happening? Where are people getting their information?” Then they need to be more agile and adept and say, “We need a completely different approach.”
OH: If everyone has access to the same AI, how can a GTM model still differentiate?
JB: There will be more competitors who use the phrase “next best action” because it is the cool AI thing. But how clever is it? That’s a different story. The way you stay ahead is deciding what types of data need to be behind it. What kind of process do you need to change? What is the timing of some of these alerts? How are we part of people’s workflows? And all of this is domain. So, we’ll have to really make domain our competitive advantage.
MW: And orchestration is the multiplier. Anyone can launch a point solution, but few can connect insights across brands, roles and channels into one seamless experience. When you do, the customer doesn’t see isolated interactions—they feel one continuous relationship. That’s hard to copy, and it’s where the biggest GTM wins come from.
One of the coolest things we’ve been talking to clients about in terms of contextual data is that all of a sudden, you can do things with data that others haven’t thought of. So, I think companies have to look past just the point—just the app, just the solution—and think about the sustained advantage they’re creating with the vendors they’re working with.
OH: Personalization can build loyalty or cause backlash. How do you make sure it’s the former?
AS: Companies should practice restraint. Just because you have data doesn’t mean you should always use all of it. Remember when Delta Airlines’ dynamic pricing drew congressional pushback for appearing to exploit customer data. Or when Wendy’s faced similar backlash for raising prices during peak hours. There are a lot of examples of companies needing to be more thoughtful.
MW: And you have to combine that restraint with genuine human insight. After a pregnancy loss, I received a diaper sample from a brand I’d registered with. There’s no way they could’ve known about this loss from the data back then, but it was a painful miss. Contrast that with a colleague’s Stitch Fix experience. When her baby was born prematurely, the stylist noticed she removed the maternity wear settings on her account and included a children’s book for her NICU stay. Think about the power of having a human in the loop. How having someone who can interpret these little bits of data and understand how to use them to create meaningful experiences that can enhance loyalty. Products like Personalize.AI™ by ZS automate the process of building context and preferences to customize the experience at an individual level, without intruding on customer privacy. I’m excited that we’ve made that kind of progress in being human in these interactions.
OH: Arun, Maria, Jaideep, thank you so much for this conversation today.
Add insights to your inbox
We’ll send you content you’ll want to read – and put to use.
