For years, our industry has done a remarkable job of collecting customer feedback. Surveys, call centre logs, online reviews, open-text comment fields — the data has piled up at both the OEM and retail levels. In many respects, that’s a testament to how seriously we take the customer experience.
The problem? Most of that data has just been sitting there.
The challenge has always been the nature of the data itself. Qualitative feedback — what customers write and say in their own words — doesn’t lend itself to traditional mathematical analysis. You can’t run a regression on it. You can’t easily put it in a dashboard.
So while the quantitative data got analyzed and acted upon, the qualitative data was largely set aside.
The best we could do was sample it. Pull a selection of verbatim customer comments to illustrate the points being made by the quantitative analysis. That’s better than nothing. But a handful of comments from a pool of thousands of responses doesn’t tell you what’s actually in that data. It tells you what you went looking for.
That has now changed.
AI tools are being applied across the automotive industry in an expanding range of applications, and some of the most exciting are purpose-built to unlock exactly this kind of qualitative data.
These tools are designed to bring genuine analytical capability to open-text feedback, and they do it in a way that addresses legitimate concerns around AI governance and liability.
Here’s something worth understanding about AI: it is a probabilistic technology, not a deterministic one.
As UK-based AI expert David Kolb has written, these systems predict the next token based on patterns in training data. There is no comprehension, no intent, no awareness. You cannot fix that. But you can design around it.
The best tools in this space are built with exactly that in mind. They provide what I’d call a popcorn trail — a visible link between the qualitative data and the quantitative conclusions the AI is drawing from it. You can see what’s driving the insight.
That transparency is what makes these tools genuinely trustworthy for business decisions.
So what does that mean in practice?
Consider DTUs, or Difficult to Understand items. That acronym has been around forever. It refers to the things customers call or write about because they can’t figure them out: pairing a Bluetooth device, navigating an infotainment system, or understanding a new vehicle feature.
In addition to publishing Canadian auto dealer, we’ve spent years helping OEMs develop instructional and training content designed to address DTUs. That work has always been valuable. But topic selection was never guided by a comprehensive view of the data. We made do with samples.
Now, for the first time, we can analyze all of it and generate a quantitative picture of the key themes and trends contained within years of customer feedback.
That changes what’s possible.
I spend a lot of time helping industry associations select topics and speakers for major events. I can tell you that AI is always on the agenda, but so is a legitimate concern that the conversation stays too theoretical.
I hear it regularly: “We need specific examples that are making a difference today, not a high-level discussion about where AI is headed.”
This is one of those examples.
We’re in an industry that sells a complicated product — one that has a real impact on people’s lives and generates genuine passion, loyalty and, in some cases, real frustration.
We’ve always collected feedback that reflects all of that.
Now, AI tools are giving us a way to actually hear it. All of it. And to understand it in a way that leads to concrete, practical action.
That’s not a theory. That’s a business opportunity sitting in your data right now.
If you’d like to explore this further, reach out. I’d love to talk about it.
