Parts Town used to differentiate itself by promising to answer phone calls in no more than three rings, so restaurant owners would know they could quickly order repair parts when equipment broke down. But now the customers expect more digital tools and self-service options, Parts Town chief information officer Jamie Head told the Applied AI for Distributors conference Thursday in Chicago.
“That’s challenging how we think about our brand and our value proposition,” he said in his presentation entitled “Wins and Learnings in Digitization and AI.”
Head detailed several wins Parts Town has registered from a slew of AI-based initiatives. But he also highlighted the risks and challenges AI brings, including the need for strong vendor management and internal controls to prevent AI-induced errors.
He pointed to reports of some large companies scaling back AI initiatives following disappointing initial results, and attendees to ask themselves, “Are we getting the value from AI? Is it doing what it said it would do? You must be so focused on the right challenges.”
Successful AI implementations
Probably more than most distributors, Parts Town can point to significant successful deployments of AI-based systems, and Head described several examples.
One is a tool called PartPredictor, introduced last year, which allows a technician to find the right part by entering the equipment manufacturer, the model number, and the issue. The AI-based system combines real data from technicians and the power of AI to find the parts most associated with that scenario. Head said an updated version under development will incorporate data from even more manufacturers and feature more models.
PartPredictor was built by the Red Lightning Group, a team devoted to developing technological innovations for Parts Town Unlimited, the parent company, which generates $2.5billion in annual revenue, and serves over 170 countries.
Parts Town also deployed an automated warehouse system from vendor AutoStore that uses AI to direct 125 warehouse robots collaborating with human employees to put inbound products into the optimal locations and to pick orders. Head said the system puts parts away four times faster than people did in the past and frees up humans to focus on other tasks.
By calculating the best locations for commonly ordered products, the system can pick 70% of orders within only 12% of the warehouse space, minimizing unnecessary travel within the distribution center. In another example, Head discussed how Parts Town, which photographs popular parts from many angles so it can show customers 360-degree views on its website, could adopt AI solutions in the future to create three-dimensional images from flat product photos provided by manufacturers.
Parts Town Unlimited’s Home Division also uses an AI tool to analyze inbound customer service calls to gain insights into customer needs, Head said. For example, the sentiment-analysis tool determined that the three most common reasons customers pick up the phone to communicate, instead of using self-service tools, are because they need fast service, want expert advice or have complicated problems.
The risks and challenges of AI
Head said Parts Town is experimenting with other uses of AI, while anticipating new challenges the technology will pose soon.
AI also is raising new safety issues, Head said. For example, bad actors can place calls to a company, claiming to be a customer or employee—and use AI to imitate that person’s voice. AI- based cybersecurity technology will be needed to spot such fakes and detect other unusual activity. “We have to use AI to fight AI,” he said.
Vendor management is another area of concern. He said companies can benefit by collaborating closely with vendors that are embedding AI into their products. But they must also be aware that in the European Union and some other jurisdictions hold client companies responsible for errors that AI-based systems make, even if the fault lies with a vendor’s technology.
Additionally, humans, he said, will play an indispensable role in preventing errors by developing AI governance structures, centralized polices and guardrails. As AI becomes more ubiquitous, Head said, “strong ownership” of those systems will be essential. “That’s why human intervention is needed: AI makes mistakes,” he said. “We know this happens.”
Don’t miss any content from Distribution Strategy Group. Join our list.
Don Davis, former editor-in-chief of Internet Retailer magazine and Vertical Web Media, is a freelance writer based in Chicago. His experience in retail and distribution goes back to his childhood when he worked in the toy wholesale business founded by his father and two uncles and in their discount department stores located throughout the New York metropolitan area.