The retail giant and the NFL team joined other organizations at the Enterprise Connect AI conference to discuss lessons learned from deploying the technology.
SANTA CLARA, Calif. -- Walmart and the San Francisco 49ers run vastly different businesses but share a critical challenge: harnessing cutting-edge AI technology to boost profits.
This week, the retail behemoth and the NFL team joined other organizations at the Enterprise Connect AI conference to discuss what they've learned from deploying a technology that's in its nascent stage. All agreed that using AI modestly to start was the best way to tap its potential to make dramatic changes to their businesses.
Turning to AI first to solve every business problem is a mistake, said David Glick, vice president of enterprise business services at Walmart, during his keynote.
"Let's not get lost in the AI hype," Glick said. "We need to be careful that we aren't looking around for nails just because we have a hammer. Focusing on the right tool for the right use case is important."
Other companies that shared their experiences with conference attendees included food service and facility management company Aramark and biotechnology company Amicus Therapeutics. Aramark, Amicus, Walmart and the 49ers planned to deploy AI gradually to give themselves time to digest the teachings of each project.
The enterprises are among the relatively small portion of organizations that have deployed generative AI, the most advanced form of AI so far. In a survey of executives and IT professionals from 832 companies globally, TechTarget's Enterprise Strategy found that only 8% said they use GenAI across their organization.
However, an additional 22% are in early production, 33% are experimenting with the technology and 14% plan to begin adoption within the next 24 months. So, many enterprises are all in on investigating generative AI's potential to remake business operations.
Glick, who found GenAI very good at document summarization, used it to help agents at Walmart's benefits help desk. His team dropped the company's 300-page benefits guide into a vector database that fed data into a GenAI model to provide quick access to benefits information.
The system replaced the sticky notes contact center agents had on their monitors to address common benefits questions from store associates, Glick said.
While the project sounds modest, it's an example of what GenAI is good for today, he said.
"What we found, after digging into this stuff, is that we believe there is a future for GenAI, but rather than having one revolutionary thing, the future is actually a lot of little things," Glick said.
Glick's team, which focuses on Walmart's finance and human resource departments, also built an AI assistant that people at Walmart headquarters use to answer job-related questions. The company plans to roll out the virtual assistant, called My Assistant, to all its stores eventually. Walmart has 5,200 stores in the United States, including its Sam's Club locations.
What people are asking the virtual assistant can reveal areas within jobs that technology could make more efficient, Glick said. "The hardest thing to do is to know what to build [next]."
San Francisco 49ers target food service, parking
The 49ers and Levi's Stadium, where the team plays, want to use AI to improve parking and food service. The partners plan to complete the improvements by 2026, when the stadium will host six World Cup games and the Super Bowl.
The 49ers are tracking fan use of two checkout-free food stands to determine whether to add others, said Costa Kladianos, executive vice president of technology at the 49ers. The team is also investigating the use of facial authentication as an alternative to tickets and paying for food by cash or credit card. The technology would cut the checkout time at concessions from 4 minutes to 30 seconds.
The team is working with Google on using AI to provide the best routes for getting from a fan's home to a parking spot using the 49ers mobile app, Kladianos said.
Recently, the 49ers launched technology predicting wait times at food stands, restrooms and gates to the stadium's parking lots, Kladianos said. The AI-powered software lets stadium managers move their 4,000 workers to places where they're needed the most. The 49ers developed the wait-time technology with Intel and SAP.
"Being able to predict the exact amount of staffing that we'll need is huge for us," Kladianos said.
In June, Aramark, which provides campus dining services for more than 275 U.S. colleges and universities, launched an AI-powered virtual assistant called SAM that it built with AI assistant developer Cognigy. The chatbot answers students' questions about dining and meal plans. The bot, available on schools' websites, also directs them to the appropriate person if additional information is needed.
The AI assistant sometimes has to deal with inappropriate questions. For example, a Florida State University student asked the bot where he could find beer. Rather than tell the student it doesn't provide that information, the AI assistant directed him to stations on campus to refill water bottles and suggested, half-jokingly, that the student find other ways to hydrate himself, said Kate Balliet, the head of experience design at Aramark.
"It's a way to react to unexpected questions that feels human," Balliet said.
The one difference we've really seen with AI, as opposed to other standard tech installations, is that it's not set it and forget it. You really have to nurture it.
Kate BallietHead of experience design, Aramark
Keeping the assistant friendly increases the likelihood that students will use it, Balliet said in an interview. "We are seeing more conversations by volume."
Companies that want to launch virtual assistants should start with a relatively simple project to provide time to test, iterate and build the underlying technology, Balliet said. Also, companies should treat any AI deployment as a long-term project.
"The one difference we've really seen with AI, as opposed to other standard tech installations, is that it's not set it and forget it," she said. "You really have to nurture it."
AI in medicine
Biotechnology company Amicus Therapeutics has a working group comprising people from across its business units that focuses on using AI to improve the development process of medicines to treat rare diseases, the company's specialty. The group is also evaluating AI for use in patient care technology.
"It's not a top-down approach," said Gary LaSasso, senior director of global IT innovation at Amicus. "It's coming from the business. We're aligned with our business strategy and drivers."
Not all AI projects work the first time, LaSasso said. The company had to relaunch a chatbot called Amy for Microsoft Teams that scanned the company's knowledge base to provide possible remedies for IT issues.
Employees didn't use the first version much because it didn't have the GenAI technology needed to search unstructured data, such as HTML documents on the company's intranet, LaSasso said. "I recommend making sure you have solid data available."
The chatbot runs on the AI-powered copilot platform developed by Moveworks.
LaSasso said he believes AI can go beyond just making Amicus' operations more efficient. He added that he expects the technology to change how the company finds treatments for rare diseases by providing better data on the effects of drugs on biological processes.
"We're just scratching the surface," he said. "When we look back in five years, we're going to have treatments and cures for things because of AI."
Kladianos said he believes the 49ers will someday use AI to evaluate players and determine who is at the highest risk for specific injuries based on their playing styles and how to minimize the danger by altering training routines.
"If we're able to predict and prevent injuries, that'll fundamentally change the game of football," he said. "That'll fundamentally change sports."
Antone Gonsalves is an editor at large for TechTarget Editorial, reporting on industry trends critical to enterprise tech buyers. He has worked in tech journalism for 25 years and is based in San Francisco. Have a news tip? Please drop him an email.
