You Don't Have a CX Problem. You Have a Priorities Problem.

Why throwing more headcount, a BPO, or AI at your customer experience isn't working, and what actually does.

It's Tuesday afternoon and you're staring at a dashboard that won't stop sliding in the wrong direction. CSAT is down three points. Response times are creeping past your SLA. Your support team is quietly burning out, and two of your best people have started "exploring new opportunities." Your board asked about AI on the last call, again. Your CFO wants to know why cost-per-ticket keeps going up even though volume is flat.

You've been here before. You threw bodies at it last time. Maybe you considered a BPO. Maybe you signed with one of those AI vendors who promised to "deflect 40% of tickets," and six months later you're back in the same conversation with a different name on the invoice.

If any of that's landing, I'm going to say something that might feel uncomfortable:

You don't have a CX problem. You have a priorities problem.

The pattern

When CX starts breaking at most growing companies, leadership feels the pressure from the board, the CFO, customers leaving reviews. So they reach for the fix of the moment. For years it was throwing bodies at the problem. Then it was "let's stand up a BPO." Now, in 2026, it's AI. Every strategy deck has it. Every vendor leads with it. Every CX leader I talk to is getting asked "what's our AI strategy" at least once a week.

None of those fixes are wrong in isolation. The problem is why they're being chosen. They're chosen because they move a number this quarter. Cost-per-ticket goes down when you outsource. Response time goes down when you add headcount. Deflection rate goes up when you launch a chatbot. Meanwhile, the thing that actually drives all of those metrics, the customer experience itself, doesn't improve. It just gets handled differently.

That's not a CX strategy. That's metric management.

Why none of it sticks

When a decision doesn't start with the customer, every shortcut downstream fails for the same reason: it's optimized for the wrong thing.

A BPO agent can close a ticket quickly, but if they don't know your product well enough to actually resolve the issue, the customer comes back two days later angrier.

An AI chatbot can deflect a ticket, but if your knowledge base is outdated, it deflects with the wrong answer and the customer escalates. More headcount can handle volume, but if the issues generating that volume aren't being diagnosed, you're paying more people to process the same broken experience.

The numbers move for a month or a quarter. Then they slide back. Then you're in the same room, asking the same question, and someone suggests the next fix.

Leaders default to quantitative data because it's what they have and what their board wants. But the data alone doesn't tell you what's broken. It tells you what's breaking. Those are different questions. What's broken lives with the people closest to the customer: the support agents, the success managers, the frontline team that talks to humans every day and knows exactly where the experience falls apart. They're almost never in the room when decisions get made. And that's the priorities problem.

What this looked like in real life

A couple of years ago, my team inherited a months-long backlog of tickets that had been escalated to Engineering and were just sitting there. One of our customer-facing teams kept sending issues up, and nothing was coming back. Members were waiting weeks for resolutions.

The obvious fix, already on the table before I got involved, was to stand up an escalation team. A middle-layer group that would process the backlog faster and triage new tickets before they hit Engineering. More bodies, faster throughput, cleaner dashboard. If this situation happened today, the conversation would be different. It wouldn't be "let's staff an escalation team." It would be "let's build an AI agent to handle these tickets." Same pattern, different wrapper.

Here's what we did instead. Before hiring anyone or buying anything, I shadowed the people who were actually escalating tickets. I sat with the team working through the backlog. I talked to those teams' managers about what they were seeing. Nobody had pulled this data together before, because nobody had thought to.

What we found: most of the "escalations" weren't real engineering issues. They were troubleshooting gaps. Members were hitting friction and frontline teams didn't know how to identify a root cause and resolve it, so they escalated. The issues were repeatable. They were solvable. They just weren't being captured anywhere.

So we built troubleshooting guides and an internal knowledge base. We brought in some intelligent support tooling. We worked with impacted customer-facing teams on an actual escalation framework so they knew what could be resolved where. And we partnered with Engineering and Product on a prioritization framework, so when real issues came through, they got handled fast.

Seventy percent of the escalation volume disappeared. Not because we processed it faster. Most of it stopped existing. Resolution times went from weeks to under 24 hours, and a couple of quarters later, the average was 2 to 8 hours. If we had built an AI agent to handle the original backlog, we would have automated the wrong work at scale. The metric might have moved. The problem wouldn't have.

The hard part

None of this is complicated. What makes it hard is that it takes longer than a quarter and doesn't produce a clean strategy slide.

The companies that commit to this work do it because they've decided, at the leadership level, that they're actually building a customer-centric business, not just managing CX metrics. Everything downstream follows from that decision. The ones that can't make that decision keep reaching for the fix of the moment, and keep ending up in the same room looking at the same dashboard.

If any of this is hitting, you probably already know which side of that decision you're on. You also probably know the pressure isn't going to let up just because the shortcut hasn't worked. The AI vendors will keep calling. The board will keep asking. The metrics will keep sliding until something underneath actually changes.

The work still starts with the customer, and with the people closest to them. That's the unsexy part most companies skip. It's also the part that makes everything else (including AI) actually work.

✌️ I'm Lupe Gonzalez, founder of CXAmplified. I work with early- and growth-stage companies whose CX is under pressure and whose leadership knows the current path isn't working. If any of this is where you are right now, I'd love to hear what you're navigating. You can reach me at lupe@cxamplifiedconsulting.com or find me on LinkedIn.