Companies Take the Next Step in AI Deployment

To be sure, AI is at the center of conversations around the world, not only at dinner parties but also on Uber rides and airplanes — and, perhaps most especially, on LinkedIn. Often, these conversations center on fantastical futures in which the technology has wiped out entire industries or even achieved sentience. But in the present, companies such as Michelin are rapidly experimenting with, validating, implementing and optimizing AI applications that provide concrete benefits such as improved productivity, efficiency and accuracy.

Jay Titus, vice president and general manager of workforce solutions at the University of Phoenix, says that many employers are still in “discovery mode” with the technology. But he notes that AI tools — and the ways companies use them — are evolving rapidly. “We’ve come from a point where employers were fearing AI to a point where they’re embracing it,” he says. “The question used to be, ‘What do we do about AI?’ That has shifted over the past year or two to, ‘What do we do with AI?’”

EXPLORE: How AI is transforming hybrid work for businesses.

How Michelin Arms Workers With AI

Rajagopal acknowledges that there is significant hype surrounding AI. However, she says, the technological advancements of large language models (LLMs) are “very real,” and those breakthroughs are penetrating other areas of AI, such as time series forecasting, that are unrelated to the generative tools that tend to dominate headlines.

In addition to adopting Microsoft Copilot, Michelin has created its own generative AI platform, dubbed Aurora. The platform, hosted on a mix of Azure and on-premises infrastructure, gives users secure access to LLMs such as ChatGPT and Llama, and employees can use it to make inquiries about internal processes and policies.

“If they have a question about how to execute a particular finance process, instead of digging through the documentation, they can just write out the question, and the answer comes back to them with specific references to our documents,” Rajagopal says. “It’s empowering because it’s secure and accurate, and it saves employees a lot of time.”

Michelin is also in the process of rolling out an AI tool called IRIS, an automated visual inspection system for the tires produced by the company. The system, which uses computer vision and AI to scan and identify defects on tires, replaces the repetitive and “ergonomically challenging” task of visual inspection, Rajagopal says. (She notes that human operators remain responsible for final quality decisions.)

The company has also implemented machine learning-based forecasting tools in its supply chain, adopted digital twin technology to more rapidly visualize tire design changes and worked with partners including Microsoft to co-develop AI solutions. Michelin is currently achieving an annual ROI of more than $50 million from its AI tools, and that number continues to grow 30% annually, Rajagopal says.

Rajagopal stresses that Michelin is looking to arm employees with new tools, not replace them with AI. “We see AI as a technology with huge potential to make work more rewarding for our people, increasing their productivity and freeing time for creativity and collaboration,” she says.

RELATED: How one company created an app to improve their customer service. 

AI Drives Supply Chain Decisions

Rather than fearing that they will be replaced by automation, many employees are eagerly embracing AI solutions, says Carl Brisco, CTO and senior vice president for the ODP Corporation. In particular, he says, employees have gravitated to the company’s Personal Assistant tool, an internal chatbot built using Microsoft OpenAI Service. The platform provides secure access to AI capabilities while ensuring data security and compliance, and the company has also created offshoots for lines of business such as HR and sales.

“I’ve had people come to me directly and thank me for building Personal Assistant,” Brisco says. “As an engineer, the best compliment you can get is when someone uses your tool and they like it. The technology is having a direct human impact that’s super positive.”

The ODP Corporation is the parent company of retailer Office Depot. The company also operates as Veyer, a supply chain and logistics business. Veyer has used AI-powered tools for years to help with route management, analysis, demand forecasting and inventory operations. Now, the business is using generative AI to help train new employees and make them more effective at their jobs. “There’s a lot of churn in a distribution center,” Brisco notes. “We’re trying to take some of the load off the trainers and the managers by giving associates direct access to answers via LLMs. They can use natural language to learn about our standard operating procedures and best practices.”

The sales bot used by ODP Business Solutions is particularly impactful, Brisco says. The tool can identify at-risk accounts and analyze margin data, but so far, sales teams are seeing the greatest efficiency gains from the bot’s SKU-matching capabilities.

Previously, when sales teams put together project bids, they would need to manually review stock-keeping unit (SKU) data line by line, taking great care to make sure their bids closely matched the scope of their competitors’. Now, the sales bot can automatically compare competitors’ product offerings to ODP’s own catalog during the bidding process, saving sales representatives four to five hours per week, Brisco says.

“Before, the process of building a bid was incredibly laborious and error-ridden,” Brisco says. “A single bid can have anywhere from 200 to 2,000 SKUs. The technology is giving people time back, and they can use that time to close more deals. Results like that have helped people overcome their initial fear that AI is going to replace workers, and now they see it as a force multiplier.” 

This shift — from viewing AI as a potential replacement for employees to seeing it as a tool to enable employees — mirrors what Titus is seeing across the business landscape. “My firm belief is that AI is only going to enhance skills,” he says. “It’s not going to replace skills. That’s the key to getting the most value from these tools, and organizations that can embrace that will get more out of their workforce.”