We may not have flying cars (yet) as predicted in Back to the Future and sci-fi from decades past, but the advent of artificial intelligence has certainly come far quicker than many may have anticipated. It’s affected many industries, including auto insurance companies by streamlining claims processing and allowing adjusters to be more productive – although this does come at a cost, too. Birthing pains, perhaps.
Tech entrepreneur Al Lijee’s US-based platform myWindshield connects Americans with local, trustworthy auto glass shops located all across the United States. While the platform is not directly associated (but tangentially is) with auto insurance, the AI technologies adopted by auto insurance providers have sparked the possibility for innovation to help facilitate communications between auto glass technicians and their customers.
Q: Al, tell us about myWindshield first and how it works.
Al Lijee: myWindshield is an online platform that helps connect users with qualified, vetted and approved auto glass shops all across the United States. We mostly work with small, independent auto glass shops trying to compete with bigger national franchises, so essentially we operate as a lead generation platform helping these smaller companies compete online with digital marketing and SEO.
The way it works from the customer’s perspective is they enter their ZIP code and are presented with a list of myWindshield member partners operating in their area. They’re free to look at their profile to see if they’d be a good fit for their needs and to call them for a free quote. Everything after that is hands-off from our perspective, and we want it that way so that the auto glass shop can operate freely and provide the best service possible without interference.
The opportunity we’re seeing now with AI, especially post-pandemic and how auto insurance companies have gone through the teething phases with it, is that we can leverage its power to add value and simplify the quote process for our auto glass member partners with their customers. It wouldn’t replace our core business model but rather be a value-added proposition.
Q: How is artificial intelligence being used in the automotive aftermarket industry, specifically for auto glass services?
Al Lijee: Well, the closest proxy we have for auto glass repair and replacement services is from auto insurance companies and how they’re using AI in recent years. I’ve heard it said that around 15% of insurance claims were settled just by photos before the pandemic and that figure has since jumped to 60% and will likely go up to 80% or beyond in the next couple of years.
Many big auto insurance providers in the US are now using AI such as Tractable and Solera’s Qapter to automate the claims process, claiming that about 25% of all claims are so on-the-nose that they don’t really need human intervention. This frees up a lot of time and resources to focus on bigger claims, for example. A lot of auto glass shops are pushing back, however, saying that the AI can be wildly inaccurate, leading to disputes where the auto glass technician is kind of stuck in the middle or even being questioned by customers coming in with inaccurate quotes. Still lots of work to do, seems like.
This poses challenges and possibilities that must be considered. We don’t want to rely heavily on the quotes from AI or endorse them, but rather we want the auto glass shop to be able to intervene and apply their expertise in cases where it’s inaccurate. If a big chunk of quotes are spot on, this tool can be very useful for customers and auto glass shops alike, saving time going back and forth sending photos of the damage to their windshield, for example.
Q: myWindshield generates relevant leads for auto glass shops, but how can you improve the process with AI solutions?
Al Lijee: The way we’re planning on using AI is with photo recognition to analyze the damage to a windshield. This particular piece of auto glass is different from side windows because windshields are usually made from laminated safety glass that can be repaired; other vehicle glass is usually tempered safety glass and usually can’t be repaired.
Essentially, the customer will upload a photo of the damage to their vehicle and the AI will determine the make and model of the vehicle (which tells the auto glass shop which piece needs to be supplied and the associated cost), where the damage is and how severe it is.
Knowing whether to repair or replace a windshield depends on many factors, usually coming down to the expertise of the auto glass technician, state laws for windshield obstructions/damage, and of course, safety. Small cracks or rock chips are usually not a big deal to repair, but cracks that extend to the edge of the glass or that directly obstruct the driver’s field of vision usually require full windshield replacement. There are many variables to consider, but whenever it’s possible to do so, windshield repair tends to be a lot cheaper than windshield replacement.
Automating the process of determining the severity of damage to the windshield can greatly improve the efficiency of auto glass technicians and with image recognition from a massive databank of different vehicles, types of damage and so on, we believe it is possible to leverage AI to give reasonably good quotes in a fraction of the time it takes for an auto glass technician to do so. Still, we firmly believe in the honest and hard work and expertise of our auto glass member partners and the technology will always come second to human experience.
Q: What are some of the key challenges to using AI for generating quotes from auto glass shops affiliated with myWindshield?
Al Lijee: So I’ve seen a lot of articles of auto glass shop owners complaining about the inaccuracy of AI-generated quotes for insurance claims, some even saying that 99.9% of all quotes are bogus and inaccurate. That may be the case and there are certainly many kinks to work out with the AI, but I think it’s likely that in the coming years it’ll only get better and better, more and more accurate.
A big challenge we want to overcome – or avoid, when possible – is using AI as a crutch for all quotes. We certainly don’t want to endorse the results of an AI as a quote written in stone that our member partners must honor. We want to make it clear that the customer’s photo will be reviewed by the auto glass shop for an accurate quote, which may or may not align with what the AI quotes.
By offering this value-added service for our users and member partners, it would be something completely optional that could potentially save lots of time and resources but could also be completely ignored if the auto glass shop is old school, say, and doesn’t mind manually quoting customers.
myWindshield is always growing and onboarding new auto glass shop member partners from all across the United States. We’re currently updating our website content and design to be more user-friendly and plan to continue adding valuable information and resources for users.