Artificial intelligence and machine learning have both been the subject of breathless, ceaseless hype in the media and C-suites around the world, but what does all of it actually mean for the automotive sector? It might seem that just mentioning these buzzwords will magically raise your valuation, but the reality is that an AI-powered solution is not innately better than a solution that doesn’t use AI. As with many other transformative changes in the industry, getting the most out of AI is a matter of pointing the right technology at the right problem – and bringing the right tools for the job is innately a human problem.
That’s not to undersell the major disruption we are seeing right now, of course. AI is almost certainly poised to be the cornerstone of the fourth industrial revolution. The third industrial revolution was about automation; this one is about intelligence – not just working or not working, but working in a specific way given certain external influences or past behaviors.
But as with all other major developments, it is important to assess what the technology is actually doing and whether it makes sense in a given context. The tire industry – and the aftermarket space in particular – is indeed a prime example of where AI can make a positive change. However, we still need to use our ‘real’ intelligence when considering how to apply our ‘artificial’ intelligence.
AI and machine learning are, ultimately, an enhanced avenue of statistics. They learn from actions and results in the past and use that to make an estimation of the future. The aftermarket is a great place to apply this, since there’s a rich data asset that is highly fragmented. People in the tire industry don’t have much wiggle room in their daily schedules, so the time savings pile up from even small efficiency gains. Pointing an AI at a set of specifications, preferences and needs for a customer and having it reliably come back with the perfect tire recommendation ends up being enormously valuable to businesses and their customers.
We mustn’t let those optimizations blind us, though; AI is far from capable of doing the job alone. Just like humans, AI can make mistakes. More concerning, it can make truly shocking mistakes that even an entry-level employee wouldn’t make – often the result of biased, incomplete, errant or otherwise flawed data. That’s why AI is best viewed as another tool in an ever-expanding toolbox. Ideally, this will be backed up by an experienced industry veteran to check the work, plus someone who understands data and the AI tool well enough to fix any issues and ensure the model becomes more reliable next time. In that sense, AI is rightfully far from taking anyone’s job away – in fact, the opposite holds true. It stands to create many more jobs if implemented correctly.
Getting the most out of these new technologies means being ever-vigilant about your data set and sharpening practices moving forward. We are in an interesting place right now, since Covid-19 caused a major disruption in consumer behavior and the supply chain, and we’ve only been able to count data as ‘normal’ again for a fairly short time. If you haven’t been diligent enough about boosting the quality of your data, now’s a great time to start.
A quality data set that is ‘AI-ready’ will have three key traits: scale, granularity and cleanliness. Scale simply refers to the breadth of information – the longer you track, the better it gets. If scale is breadth, granularity is depth – for example, breaking out one price line into five or six different ones. As breadth and depth expand, cleanliness becomes ever more important – this refers to correcting inputs that were entered incorrectly, whether due to a simple slip of the finger, a misunderstanding or carelessness. Major culprits for warped data in the tire industry are whether a price field includes tax or not, and whether it includes installation, a tire disposal fee or other variables that are easily missed.
This sounds like a lot to take in, but it will be well worth businesses in the tire space getting their heads around AI soon. Having rich, well-kept data will enable AI and machine learning algorithms to be easily layered onto the existing business in an impactful way. Coupled with a strong, capable human team, this will naturally put businesses that understand their tools well ahead of those who skip AI entirely or – potentially much worse – blindly buy into the hype and expect it to solve all their problems. We are at a major turning point in many industries right now, and the automotive sector can’t afford to waste time spinning its wheels.