I want to tell you something I learned sitting in a CFO seat.
Numbers do not lie. But they are always late.
By the time a problem shows up in the financial data, it has been living in the organization for months. You see the decline and think: what changed? But the change happened earlier. It happened in a meeting where nobody had clarity about who owned a decision. It happened in a quarter where a team kept moving without stopping to look at what the work was actually producing.
I have been watching what happens to organizations when artificial intelligence (AI) enters that environment. This article explains what I’m seeing.
The investment is going up. the results are going down
This is not an opinion. Forrester's 2024 data showed that US customer experience quality has fallen to its lowest point since 2016. At the same time, CX investment was at an all-time high.
More investment. Worse results.
As many as 39 percent of brands experienced significant CX declines last year. More than double the year before. And only three percent of US brands are genuinely customer-focused, despite 90 percent claiming it as a core strategy.
Something is deeply wrong with how organizations are making decisions about this. The organizations advancing are not the ones adding the most technology. They are the ones that stopped and asked the harder question first: what are we actually trying to make possible for people?
AI did not create this. it made it impossible to hide
There is a reason 77 percent of customers experienced a service problem last year. And 64 percent described what they felt as rage.
The research is clear about what drove it. Not bad service. Blocked access to service. Automation built to deflect. Systems designed to reduce what a contact costs, with no attention to what that contact means to the person making it.
That is what artificial leadership looks like. It moves fast. It scales. And it creates exactly the kind of experience the data is reporting.
Organizations that deployed AI to reduce cost-per-contact, without asking what kind of effort they were creating, are now living in that number.
Satisfaction tells you what happened. trust tells you what comes next
These are two different measurements. Most organizations are running one of them and calling it enough.
Satisfaction measures what customers report after an interaction. It is a lagging signal. By the time a satisfaction decline shows up consistently in your data, the trust that drives purchasing behavior has already moved.
Trust measures predictability. Competence. Whether people believe you will do what you say. And it shows up in financial behavior long before it shows up in survey scores.
I have read balance sheets where satisfaction held steady while trust was eroding underneath it. The signal appears in retention first. Then in acquisition cost. Then in the quarterly numbers leadership is suddenly trying to explain.
The people behind the technology are the variable
The organizations achieving real results with AI are not those with the best platforms. They are those where, before any technology was deployed, leadership had answered a specific question: do the people in this organization know what they own, what it requires, and what they are accountable for delivering?
When the answer is yes, AI extends that capability. When the answer is anything else, AI makes the absence of it visible to every customer who encounters it.
Employee experience is not separate from customer experience. It is the foundation of it. Employees who trust their organization transmit that trust to customers. Distrust is also transmitted.
When you deploy AI into an environment where frontline employees are unclear, unsupported, or disconnected from what they are building, the AI inherits that environment. That is a leadership decision and it is made before any technology is procured.
Quick links
- CX is chasing ROI – but still thinking reactively
- AI and customer trust: Reflections from CCW 2026 Sydney
- The top 50 AI leaders in CX to follow in 2026