St. Louis company says AI will change business

OpenAI had just released ChatGPT, and it was clear that artificial intelligence technology had taken a historic leap forward. Employees at IT and high-tech companies began experimenting with the program while on vacation, and when they returned to the office, they all seemed to be talking about his ChatGPT, said VP of Advanced Technology Center at Worldwide Technology in Maryland. Tim Denny recalls. Heights.

Aaron Frydenberg, the center's director, asked ChatGPT for statistics on St. Louis Cardinals legend Albert Pujols.

“It can spit out answers right away,” he said. “It was really quick.”

The program was easy to use and could be used by anyone, not just experts, for experiments. This has captured the imagination of almost everyone in the technology industry, and generative AI has risen to the top of the priority list among World Wide customers, primarily large Fortune 1000 companies.

“Everyone was playing around with things on their own. And I think that's when hearts started to open up and say, 'Okay, the future is different,'” Denny said.

Over the next year, companies across industries in the St. Louis region explored how generative AI, which can create new data, images, and text based on what it learns, could make their businesses more efficient and more competitive. .

Leaders are considering the benefits and risks of technology.

In December, World Wide Technology announced plans to invest $500 million over the next three years to accelerate the adoption of AI among customers.

From technology companies like World Wide Technology to hospitals, manufacturers, investment firms, and agricultural suppliers, all companies can save time on menial tasks for their employees and increase the information available to employees and customers at their fingertips. I'm starting to think of a way.

Emerson, the Ferguson-based factory automation giant with customers and employees in most countries around the world, wants to use generative AI for translation, said Clint, Emerson's director of digital services, cloud and AI. Schneider said. The company may have employees in California who are experts in specific brands of control valves (an expertise only a dozen or so people have worldwide) and serve customers in the United States as well as South Korea. There is.

And for newly hired salespeople, Schneider said generative AI can answer questions that help them better understand a product line, and can do so more intelligently than any search engine.

Meanwhile, experts from Bayer's Crop Science division have started working on a prototype generative AI program for farmers.

Bayer is already making recommendations to farmers about decisions such as planting, irrigation, harvesting, timing of fertilizer application and the density at which fields are seeded, all tailored to the farm, environmental conditions and weather. But Nitin Nahata, senior engineering director at Bayer's digital agriculture unit Climate, said AI makes it easier to understand and question that kind of data.

“Humans can now easily ask very intuitive questions, such as 'What impact did a particular variable have on last year's revenue?' Only through a chatbot-type interface,” Nabata said. “For example, 'Hey, show me a field that lacks irrigation,' or 'Show me a field that rained last night.' …'Can you tell me how planting date affects wheat yield? .”

Bayer is testing a prototype, but it's still in the early stages, Nahata said.

Nalini Polavarapu, vice president of enterprise analytics and data science at Bayer, also sees the potential for AI to help crop scientists who need to answer a wide range of questions from growers. .

“The range of topics is huge,” says Polavarapu. “No one can be better at anything.”

Previously, if there was a problem the company wanted to address, the solution might have come from five Bayer leaders sitting in a room, she said. Now, her group of customers and businesses are hearing about her ChatGPT, imagining ways to use such technology, and approaching the company about potential uses.

“It opens up opportunities to a lot of people with domain knowledge and really democratizes the use of AI,” says Polavarapu.

World Wide Technology's Frydenberg said one thing is certain. That means executives need to start considering how generative AI will impact their business. Because your employees are likely already using it.

There are legal and ethical issues to overcome, from how intellectual property is used to the risk of creating bad data, also known as AI 'hallucinations'.

“If you do it right, there are great benefits. If you do it wrong, there can be great liability and risk,” said Chad Bockert, vice president of marketing for Worldwide Technologies.

At Emerson, leaders told employees early on that ChatGPT learns through human feedback, so any information entered into ChatGPT can be used to train the model. To protect Emerson's intellectual property, the company asked employees not to use the program for business purposes, said Schneider, director of digital services.

“We take a zero trust approach to AI and generative AI tools,” Schneider said.

The company has established a governance team for its use of AI. The team includes Mr. Schneider and members of his IT, legal and cybersecurity departments at Emerson to discuss the technology's potential uses and possible unintended consequences.

Other companies are similarly cautious.

Chesterfield-based Mercy Hospital System began exploring potential uses for AI across its health system last year, and in December released a chatbot to answer patient questions online. Last year, an executive said Mercy was developing an internal AI “code of conduct.”

World Wide Technology staff said they provide instructional videos, articles and quizzes for employees who are new to the technology. We call this process “AI driving license.”

But with these safeguards in place, companies are starting to imagine how generative AI will change the way they do business.

Companies are investing now in the belief that AI will become a critical tool for efficiency and decision-making in the not-too-distant future. They want AI to save employees time on menial tasks, or tap into mountains of unorganized data that are created during normal work hours but are too large or unorganized for human time to investigate. I imagine doing that.

“We see this as a huge opportunity,” said Bockert, head of worldwide marketing. “We're investing to be at the forefront of being able to help large organizations. …It's a bit of a gold rush.”

“It's just the way people's imaginations guide them,” Bockert continued. “It significantly reduces a very manual and time-consuming procedure…from hours to minutes.”

Of course, change doesn't happen overnight.

Kirk McDonald, a portfolio manager and senior research analyst at Clayton’s boutique investment firm, began experimenting with AI-based tools around 2017.

His firm, Argent Capital Management, started with a program that analyzed what executives at public companies said on quarterly conference calls with investors. It provided an overall score, suggested whether the tone of the call was positive or negative, and then provided separate scores for different sections of the call.

Was the tone of the CEO positive or negative? What about the CFO? During the Q&A portion, were the questions from the financial analysts positive or negative? For example, a phrase like “Revenue growth struggled in the quarter.” That's a clear red flag. MacDonald said Argent has incorporated the software into its stock price performance prediction model, but said its significance is very small.

After about a year, he realized it wasn't adding much value anymore.

“I didn't really understand why,” McDonald said. “The problem could not be identified.”

He wasn't sure if the technology evolved due to changes in the company that manufactured it. Or maybe his company just takes a longer-term view when it comes to investing than software, he wondered. Nevertheless, Argent stopped using it.

The company also tried another program to analyze financial reports. That was impressive in a way, but equally I found it didn't add much value.

“I was using it at a time when interest rates were going up, so it didn't really matter what was said on the phone,” said senior research analyst Bill Weeks.

They tried a third piece of software that provided analysts with an overview of financial results. But it turns out important parts weren't being captured, McDonald said.

“It wasn't reliable yet,” he says.

Over the next few years, he believes this will change the way investment analysts work and make them better at their jobs. Better tools for analyzing earnings reports are important if they are to perform well enough for investors to rely on them. And having a program that can summarize those calls and get to the important parts can save you a lot of time.

“We're going to keep trying until we find a way to use it,” McDonald said.

Still, neither McDonald nor Weeks believe that AI-based tools can replace human work.

“Human judgment is very important,” Weeks said. “At the end of the day, we need to know what these companies are doing.”