How are companies using AI agents? Here’s a look at five early users of the bots

These autonomous AI agents can follow instructions and do things from checking a car rental reservation at the airport to screening potential sales leads.

Software companies from Salesforce to ServiceNow, Microsoft and Workday last year all announced their own AI agents, which they say can help businesses be even more hands-off in areas like recruiting employees, contacting potential sales leads, creating marketing content and managing their information-technology.

If these AI agents work as promised, they could also provide businesses with the return on investment they have been looking for out of generative AI. According to some corporate technology leaders, that means the ability to tie the technology to a reduction in the number of hours employees work, or even how many new people they need to hire.

Phu Nguyen, the head of digital workplace at Pure Storage, considers AI agents an obvious boost for each of the data storage firm’s employees: “Why should executives be the only people that have a ghost writer that writes their emails or does their slides? Imagine, now, all employees have that power?" he said.

Still, more AI agents can mean more problems, especially in cybersecurity, according to market research and IT consulting firm Gartner. By 2028, at least 15% of daily business decisions will be made autonomously through agentic AI—up from 0% in 2024, Gartner said. But, also by that time, 25% of enterprise breaches will be tied to AI agent abuse.

Here are five companies that have started integrating AI agents into their products and operations, and what they have learned in the process.

Johnson & Johnson: Drug discovery agents

At New Jersey-based Johnson & Johnson, AI agents are being used to help the healthcare giant with the chemical synthesis process in drug discovery.

Once a promising pharmaceutical molecule has been identified, according to Chief Information Officer Jim Swanson, it needs to be measured for its cost effectiveness and reliability. And there are a lot of variables to consider in doing so—from temperature to which reactions are being optimized.

The solution: an autonomous AI agent that can determine the best time to conduct a solvent switch, a process where one solvent is swapped for another to crystallize a molecule and actually create the drug, Swanson said.

Without the help of these technologies, J&J’s scientists would go through multiple iterations of the same process manually, ensuring the right conditions are in place to optimize the switch.

“We’re using agents now to take that content, with all those variables, and figure out, ‘When’s the next time to do that switch and actually execute it?’" he said. The agents are combined with traditional machine learning and digital twins—essentially digital replicas of real-world entities—to speed up the process, he said.

Still, J&J is proceeding with caution. It is being “mindful of this risk with autonomous agents that could be creating bad behaviors," Swanson said, such as generating biased information and outputting hallucinations.

Employees review the outputs of its agents, he said, but the company is still figuring out how that oversight can be done more systematically.

Moody’s: Financial analysis agents that disagree

AI agents are becoming key players in research at Moody’s, the New York-based financial analysis and software company.

Many aspects of research, including industry comparisons and looking at companies’ Securities and Exchange Commission filings, were already outsourced to lower-cost areas outside the U.S., said Nick Reed, the company’s chief product officer. Now, some of that work is being done by autonomous AI agents—specifically those that work in conjunction with other agents.

The company has developed a total of 35 agents, some for smaller tasks like project management, and linked them up with agents for supervising them, creating what Reed calls a “multi-agent system."

Moody’s agents are given specific instructions, personalities and access to data and research. As a result, they can come to different conclusions, especially for complex topics like analyzing the financial fitness of a company that appears healthy, but is facing geopolitical risk.

“It’s almost a bit like your ability as an individual person," Reed said. “What we worked out is that an agent is better at not multitasking."

EBay: Agents that write code and sell items

EBay is using AI agents to help write code and create marketing campaigns. The company also plans to roll out agents that can help buyers find items and sellers list goods.

To make it all happen, the San Jose, Calif.-based online marketplace created its own “agent framework" that can use several large language models in the background, said Nitzan Mekel-Bobrov, eBay’s chief AI officer.

The agent framework functions as an orchestrator, dictating which AI models will be used for certain tasks like translating code and suggesting code snippets, Mekel-Bobrov said. As its agents become more sophisticated, they’ll be able to act more autonomously—writing more code on their own, line by line, as human developers would, he added.

“As employees interact more and more with the systems, it also learns their specific preferences," he said.

Deutsche Telekom: An ask me anything agent for employees

Telecommunications giant Deutsche Telekom, which has roughly 80,000 employees in Germany, has rolled out an AI agent for its employees to ask any question about internal policies and benefits, and for its service staff to ask questions about its products and services.

About 10,000 employees use the AI agent each week, said Jonathan Abrahamson, Deutsche Telekom’s Chief Product and Digital Officer.

Deutsche Telekom is also experimenting with letting the agent, which it calls askT, perform tasks on behalf of employees, Abrahamson added. An employee who wants to book their next vacation, for instance, can tell askT to put the leave request into the human resources software system on their behalf.

Cosentino: ‘Digital workforce’ solving customer issues

The Spanish company Cosentino, which makes countertop surfaces and other stone materials for homes and buildings, has brought on a “digital workforce" of AI agents to fill the gaps in its customer service staff, said Rafael Domene, the company’s CIO.

Cosentino calls its agents digital workers because they’re treated that way—they need to come with a set of basic skills, but they also get training when they first arrive on the job. They’re also instructed to follow a strict process, and the company knows if they go off the rails, Domene said.

Now, its “digital staff" have entirely replaced the work of three to four people previously involved in clearing customer orders, and those staff are focusing on other areas of service, according to Domene.

Write to Belle Lin at belle.lin@wsj.com