Last month we shared an interesting story on our Facebook page about a new challenge in our industry: Drivers “diagnosing” their car with ChatGPT, then asking a service centre to perform a repair based on that diagnosis. The TLDR from that post:

• This creates a difficult situation when the AI diagnosis is wrong.
• The AI diagnosis is wrong more often than it is right.
• Most AI chatbots present their responses with high confidence, often leading users to develop unwarranted trust in the information.

Now, we’ve experienced a different, but related challenge: Vehicle owners researching their vehicles issues and then asking us to carry out a test plan recommended by AI. Admittedly, this is better than just throwing parts at the problem, but it’s often still a total waste of time.

Here’s an example.
Two weeks ago, a new client contacted us about a problem with their vehicle and sent over a series of tests that they requested we perform; a test plan clearly written by AI. We advised that based on their description of the problem, most of those tests wouldn’t be necessary – or helpful – when diagnosing the issue. Instead, we suggested having one of our technicians diagnose the concern our way, using a logic-based but realistic approach where the results of each test determine which test comes next. The client declined (which is partially on us; more on that later) and stated that they felt more comfortable performing their test plan instead. “What would it cost to do it my way?”

Looking through their list of tests and measurements, we advised that it would take a technician about two hours to complete client’s request. This would result in a total cost of just under $460 at our current diagnostic rate of $219/hour, including GST. Again, we advised that many of these tests seemed unnecessary and suggested a different path forward. The client again rejected the suggestion, asking to make an appointment for the testing – and just the testing, because they would handle the repairs later. It was clear that as a newcomer who hadn’t gotten to know our team yet, they trusted AI more than they trusted us.

We performed all of the tests, laying out the results in great detail exactly as the client had requested – even including a beautiful report with photos and videos documenting every step. The whole process felt uncomfortable, but we had tried twice to dissuade the client. If they were happy to pay, we were going to deliver the best results possible. Following testing, the vehicle received all of the My Garage touches including a complimentary car wash and a couple of our branded water bottles in the cupholders. The client was very impressed, but not enough to trust us yet. We made some recommendations but they advised that they would prefer to review the results on their own, surely planning to consult with ChatGPT (or Grok, Gemini, etc.) to arrive at their “answer”.

Just over a week later, our team diagnosed the exact same issue on a very similar vehicle for a different client who agreed to let us take the reins.

We charged $95.

The moral of this story?

Find a good service centre and trust them! We understand that consumer trust in automotive service professionals is often low – and in a very small percentage of cases, for good reason. But please understand that the overwhelming majority of technicians genuinely want what’s best for you. They are good people who do this job for the love of the craft: for the satisfaction of solving problems, easing your stress, and taking something broken and making it work again.

Through the years, there are real savings to be gained by building a long-term relationship with a trusted service centre like ours. People who try to save money by tackling repairs themselves often end up paying more in the long run – taking expensive guesses, buying parts that weren’t actually needed, or fixing the same problem twice when a lower-quality repair doesn’t hold up. It’s a situation that brings to mind the old saying: “The poor man always pays twice.”

Why part of this is our fault.

As an industry, we are partially to blame for so many drivers’ unwillingness to pay for diagnosis and testing time. We haven’t done a great job showing our clients the value of these charges; showing them how much goes into a proper diagnosis. More recently, we’ve started explaining diagnostic charges like this:

Ms. Jones, here’s how I’d like to troubleshoot your concern. To start, we’ll need a skilled and caring technician. Experienced technicians don’t come cheap, but they’re worth it. We’ll equip that technician with tens of thousands of dollars’ worth of testing and diagnostic equipment, along with access to your manufacturer’s diagnostic software, which costs about $4,000 per year in subscription fees. They’ll also rely on wiring diagrams, specifications, and other service information, which is another separate subscription. Finally, they’ll need a hoist and a properly equipped facility to work in, with all of the associated costs. The good news? We can rent you all of this for just $219 per hour.

Without a proper explanation, some clients may see the diagnostic fees as a blank cheque; not understanding exactly what they are receiving in exchange. In the case of the client discussed earlier, they saw value in the ChatGPT test plan because they understood exactly what they were getting. This is a reminder to everyone in our industry: While we might do this job every day and fully understand the process, these processes must be better communicated to others.

Some thoughts on trust:

For decades in our industry, we’ve been amused by how often people will trust a piece of machinery more readily than another human being. That’s a big reason why diagnostic tools like battery testers and refrigerant analyzers often come equipped with printers these days. Even in our shop, where we’ve built a reputation for transparency, we’ve sometimes seen a client raise an eyebrow when we say their battery is testing weak. But the moment we hand them the printout, the conversation changes. In a way, today’s tendency for people to trust AI over the advice of a technician is just the latest version of the same phenomenon.

A service centre like ours is also more likely to encounter these AI-leaning clients because of our focus on electric vehicles. While this is beginning to change as EVs become more mainstream, early EV adopters have traditionally been more tech-oriented individuals – often highly intelligent but sometimes less comfortable socially. They tend to understand their vehicles better than the average car owner and often enjoy taking a more active role in its maintenance.

Closing

Aside from the energy-related and environmental concerns, there’s nothing wrong with AI. We actually use it fairly regularly within our business. We just need to understand its limitations, and remember the benefits that good, old-fashioned human interaction can offer.