For most of the last decade, the dominant story about customer experience has been a story about friction. Friction was the enemy. Long hold times, repetitive forms, sluggish processes, and the indignity of being passed from one department to another. The promise of automation, and then AI, was that friction could be engineered out of the system.
There is a moment most people have had in the last year that no one talks about in strategy meetings. You’re trying to resolve something with a company. A canceled flight, a denied claim, a charge you don’t recognize. You open the chat window. The bot greets you by name and asks how it can help. You type. It misreads you. You rephrase. It misreads you again. You type the word agent and it offers you a help article. You type it again, and it asks if you’d like to start over. By the time a person finally appears, you’re not a customer anymore. You’re an opponent.
This’s the part that the dashboards don’t capture, because the dashboards were built to capture something else.
For most of the last decade, the dominant story about customer experience has been a story about friction. Friction was the enemy. Long hold times, repetitive forms, sluggish processes, and the indignity of being passed from one department to another. The promise of automation, and then AI, was that friction could be engineered out of the system. Faster resolution. Lower cost to serve. Better experience, in theory, for everyone. The story wasn’t wrong. It was just incomplete.
Because as the operational friction came down, something else came up in its place. Something the metrics weren’t designed to see. Customers started reporting that the experience felt worse even when it was, by every measurable definition, faster. The interaction was more efficient and somehow more humiliating. The system worked, and yet people walked away feeling worked over. The category had solved one problem and unintentionally created another, and the new one’s harder to fix because it doesn’t show up in the same dashboard. The new problem isn’t friction. It’s conflict.
Friction Was Operational. Conflict Is Emotional.
The traditional definition of friction is logistical. It’s the time, the steps, and the effort it takes to get something done. Friction in that sense is annoying but impersonal. You wait, you repeat yourself, you fill out the form again, and eventually the thing happens. The cost is measured in minutes and patience.
What’s happening now is different in kind, not just in degree. Customers aren’t just frustrated by how long it takes to reach a resolution. They’re frustrated by the sense that no one inside the system is willing to recognize them. The chatbot doesn’t understand the nuance of their situation, and there’s no clear path to someone who will. The denial letter arrives without explanation, and the appeal form leads to another denial letter. The interface is fast, and the experience is cold. The faster it goes, the more it can feel like the company has decided the customer isn’t worth a human moment.
That shift, from operational annoyance to emotional injury, is what turns friction into conflict.
People will tolerate a great deal of inconvenience if they feel seen inside it. They’ll tolerate remarkably little efficiency if they feel dismissed by it. This isn’t a soft observation. It’s the underlying physics of why two interactions with the same resolution time can produce wildly different reviews, complaints, and churn outcomes. The metric is identical. The relationship isn’t.
And the more an organization automates the front line of its customer experience, the more often it ends up on the wrong side of this equation without realizing it.
The Efficiency Trap
The reason this is hard to catch from the inside is that organizations measure what they can measure, and the things that are easy to measure are the things that have already been optimized.
Call deflection rates. Average handle time. Self-service completion. Cost per ticket. Headcount per thousand customers. These are the numbers that move when AI and automation are introduced, and they almost always move in the right direction. The deck looks good. The board is pleased. The savings are real.
What doesn’t show up on the same deck is the emotional outcome of the interaction. Whether the customer felt respected. Whether they felt the company took their situation seriously. Whether they believed the process was fair. Whether they’d choose the same company again if a competitor offered the same product with a slightly better experience.
These are harder to measure, so they tend to get inferred from proxies. CSAT scores, which are often gamed or unreturned. NPS, which captures intent without much texture. Sentiment analysis on transcripts, which can flag tone but can’t read the long-term consequence of feeling unheard.
So a company can post a quarter of falling support costs and rising deflection rates while simultaneously eroding the thing that brought customers in the first place. The operational metrics improve, and the relationship metrics quietly decay, and the gap between the two doesn’t surface until a competitor lands or a public incident exposes it.
There’s a sentence worth pinning to the wall in any executive room where this conversation happens: customers can tell the difference between convenience and avoidance. They know when a company has built systems to help them faster, and they know when a company has built systems to avoid them. The interface looks the same. The intent doesn’t, and people read intent more accurately than most strategy teams give them credit for.
When Systems Feel Faceless, Conflict Escalates Faster
The other thing that’s changed is how quickly emotional friction now travels.
Inside a faceless system, accountability disappears by design. There’s no one specific to talk to, no one specific to be upset with, no one whose name the customer can hold onto. This is sometimes presented as a feature because it protects employees from the worst of customer frustration. But what it actually does is concentrate the frustration somewhere else. If no one inside the company seems accountable, the customer starts looking for accountability outside the company. The complaint moves from the call center to LinkedIn, from the help ticket to X, from the private moment of frustration to a public post that may reach more people than the company’s marketing did that week.
This is the part where minor friction becomes reputational risk. A single denied claim is a customer service issue. A denied claim accompanied by three rounds of automated responses and no path to a human becomes a story, and the story travels.
It happens most visibly in industries where the stakes are personal. Airlines, where a disruption sits on top of a missed funeral or a wrecked vacation. Banks, where a frozen account sits on top of a payroll deadline. Insurance, where a denial sits on top of a diagnosis. Healthcare, where a billing error sits on top of a treatment that was already terrifying. Retail, where a return becomes the thing that decides whether someone ever buys from the brand again.
In all of these cases, the original event is the friction. The automated handling of it is where the conflict begins. The customer isn’t asking for a faster system. They’re asking for the system to recognize that the moment they’re in isn’t a transaction.
When a company answers a moment of human stake with a workflow, the workflow becomes the story.
What The Companies Getting It Right Are Actually Doing
It’d be easy to read all of this as an argument against automation. It isn’t. The companies handling this best aren’t the ones that resisted AI. They’re the ones who thought harder about where AI belongs. The pattern that keeps appearing in organizations that are getting it right looks like this.
Automation handles the volume of work. AI handles the pattern recognition, the prep, the routing, and the drafting. And then a human shows up at the exact point where the stakes get personal, with full context already in hand, ready to do the thing that only a person can do, which is to take the situation seriously.
The work the AI does in the background is invisible to the customer. The work the person does in the foreground is the experience the customer remembers. The two together produce something neither could produce alone. The AI lets humans spend their time on the things that require judgment instead of triage. The human gives the AI’s efficiency a face that the customer can actually trust.
This isn’t a softer use of AI. It’s a more strategic one. It treats the technology as a force multiplier for the things humans are uniquely good at, rather than a substitute for them. The cost savings are still there. They’re just being earned without paying the relationship tax that pure-automation strategies quietly accept.
The companies that understand this are also designing escalation paths on purpose, not as a fallback. They make the path to a human visible, not buried. They give their human agents context, authority, and discretion, so that when a customer finally reaches one, the conversation moves forward instead of starting over. They build their automation to know what it doesn’t know, and to hand off cleanly when it reaches the edge of its competence.
It looks, from the outside, like the kind of customer experience that doesn’t need to advertise.
The customers do that for them.
The Real Frontier
Most categories spent the last several years optimizing for frictionless. The next several years are going to belong to the companies that figure out what to optimize for instead.
Frictionless isn’t the goal it sounded like. It’s a useful objective when it’s in the service of something larger, and a misleading one when it’s treated as the destination. A frictionless system that leaves the customer feeling like a number is a system that has won on the wrong scoreboard.
The goal worth designing for isn’t the absence of friction. It’s the presence of judgment. The system should be efficient, where efficiency is what the customer wants, and human, where humanity is what the customer needs, and the company should be able to tell the difference.
That’s harder to build than a faster chatbot. It requires honest internal conversations about which moments in the customer journey carry emotional weight, and which don’t. It requires a willingness to staff the high-stakes moments more generously than the cost dashboard would prefer. It requires leadership that can hold the line when the savings argument shows up, because the savings argument will always show up, and it’ll always be persuasive on its own terms.
But the companies that do this are building something that compounds. Every time a customer reaches the human moment and feels recognized inside it, the brand earns a piece of trust that no marketing budget can buy. Every time a competitor automates that same moment away, the gap widens.
At ThoughtLab, this is the frame we keep returning to with leaders rethinking their customer experience: the question isn’t how much of the journey can be automated, but which moments of it deserve to remain human. The answer is rarely obvious from the org chart, and almost never visible from the savings dashboard. It surfaces in the gap between what the company is optimizing for and what the customer is actually experiencing.
The Takeaway
AI and automation are reshaping customer experience at a speed most organizations are still catching up to. But efficiency on its own doesn’t create loyalty, and it’s never created trust. The companies that confuse the two are going to spend the next several years discovering the difference the hard way, one public incident and one quiet churn cohort at a time.
The organizations that’ll lead in the next era of customer experience won’t be the ones that automated the most processes. They’ll be the ones that understood where human interaction still matters, where empathy can’t be outsourced, and where trust is built through responsiveness, accountability, and the willingness to be present in a moment that asks for presence.
In the age of AI, the most valuable experiences won’t be the ones that feel the fastest. They’ll be the ones that still feel like someone, somewhere, was actually paying attention.
That’s the experience worth building toward. Not frictionless. Faceful.
