I have queried ChatGPT about common trouble codes, step-by-step instructions on resolving certain trouble codes, brand information for replacing various part types — and yes, even asked about its own use cases for the aftermarket.
The results varied from highly specific to uselessly generic. For instance, it showed detailed information about how to diagnose and troubleshoot a specific trouble code on a 2015 Ford F-150. It could name the best brands for replacement brake pads — with helpful descriptions — on a late-model Audi A4. On the flipside, it came up empty when asked to produce cross-references for part numbers.
Agreed, the last example may be too specific. But the essence of the aftermarket lies in its product complexity. E-commerce, while extremely successful in other industries, keeps stumbling on the auto parts hill. ChatGPT, as it stands today, may only be demonstrative. But its ultimate success in the aftermarket will depend on how well it can navigate the intricacies of vehicles, parts, catalogues and repair content.
Can AI tools recommend the right part with over 95% accuracy (in an industry where one out of every five parts is returned to the seller) to a consumer or technician? Can it consistently hit the mark on parts and inventory demands? I doubt it — and definitely not anytime soon.
The fault may not lie entirely with the technology. Data in the aftermarket continues to remain spotty, which will ultimately influence how these tools get trained. I think the real AI breakthrough for the parts industry will be in the radical improvement of data through suggestive and predictive functions. But that will take time — and significant investment. How long is anyone’s guess but I think we will probably see some breakthroughs by the end of the decade.
Till then, generative AI’s influence and contribution to our industry will be additive, not disruptive. Here are some ways generative AI could help the industry:
In the car
Automakers such as General Motors and Mercedes-Benz have announced plans to bring ChatGPT to their connected vehicle systems. The German automaker will test out the tool first with the hope of improving contextual understanding and output complexity of voice assistants. I guess the initial goal probably doesn’t go further than making the process of speaking to your car less frustrating than sitting in Toronto-area traffic.
But it would probably go further down the road, specifically for repair services. Generative AI could help drivers troubleshoot minor issues, eliminating the need for flipping through dense manuals. It could help car owners find the right repair shop and products through an interactive process.
ChatGPT could be a great tool for enthusiasts and tinkerers, helping them take care of some basic installations — almost working as a virtual technician. Depending on how ‘right to repair’ evolves, these services could open new revenue streams for both OEMs and the aftermarket.
In the store
Let’s face it: Companies have struggled with hiring and training parts store counterpeople, especially post-pandemic. Finding the right people is only going to get tougher. ChatGPT plug-ins for inventory lookup and transaction systems could help employees find the right solutions and products for customers quickly and accurately.
The technology could also help with dynamic scripting for customer support personnel. It could produce live prompts for chats and phone calls, based on what a real customer is saying.
In the shop
Generative AI could contribute the most here, at least in the short term. The most obvious use for a ChatGPT tool would be as a virtual shop assistant. For instance, it could provide step-by-step instructions on complex repairs as a technician is working on them. It could keep installers up to date on the latest repair protocols, squeezing in bite-sized training sessions during down times. Generative AI could pair with other tools such as virtual and augmented reality to provide more immersive lessons.
There are other less obvious applications as well. AI tools could assist shop owners and managers in drafting or automating customer communication. Shops are barely able to keep up with functional outreach such as onboarding or even letting customers know when their vehicle is ready. ChatGPT could help them in being more pro-active, such as documenting a detailed analysis of work done on customer vehicles, writing a welcome or thank you email, or asking them for feedback.
Kumar Saha is Toronto-based vice president (U.S.)/Managing Director (Canada) of global automotive intelligence firm Eucon. He has been advising the North American automotive industry for over a decade and is a frequent conference speaker and media commentator.
This story originally appeared in the July issue of Jobber News