Earlier this year, Mark Austin, AT&T’s vice president of data science, noticed that some of the company’s developers had started using the ChatGPT chatbot at work. When the developers got stuck, they asked ChatGPT to explain, fix, or refine their code.
It seemed groundbreaking, Mr. Austin said. However, since ChatGPT is a publicly available tool, he wondered if it was safe for businesses to use.
That’s why in January AT&T tried a product from Microsoft called Azure OpenAI Services that allows companies to build their own AI-powered chatbots. AT&T has thus created a proprietary AI assistant, Ask AT&T, to help its developers automate their coding process. AT&T customer service reps also began using the chatbot to help summarize their calls, among other things.
“Once they see what it can do, they love it,” Mr. Austin said. Forms that would previously have taken hours to fill out took just two minutes with Ask AT&T, freeing workers to focus on more complicated tasks, he said, and developers using the chatbot increased their productivity by 20 to 50 percent.
AT&T is one of many companies scrambling to find ways to harness the power of generative artificial intelligence, the technology that powers chatbots and has been buzzing Silicon Valley in recent months. Generative AI can create its own text, photos, and videos in response to prompts. These features can help automate tasks like taking meeting minutes and reduce paperwork.
To meet this new demand, technology companies are scrambling to introduce enterprise-grade products that integrate generative AI. Over the past three months, Amazon, Box, and Cisco have unveiled plans for generative AI-based products that create code, analyze documents, and summarize meetings. Salesforce also recently introduced generative AI products used in sales, marketing and its messaging service Slack, while Oracle announced a new AI capability for HR teams.
These companies are also investing more in AI development. In May, Oracle and Salesforce Ventures, Salesforce’s venture capital arm, invested in Cohere, a Toronto-based start-up focused on generative AI for business use. Oracle also resells Cohere’s technology.
“I think this is a complete breakthrough in enterprise software,” Box CEO Aaron Levie said of generative AI. He called it “this incredibly exciting opportunity where, for the first time ever, you can start to actually understand what’s inside your data in a way that wasn’t possible before.”
Many of these tech companies are following Microsoft, which has invested $13 billion in OpenAI, the maker of ChatGPT. In January, Microsoft made the Azure OpenAI Service available to customers, who can then access OpenAI’s technology to create their own versions of ChatGPT. As of May, the service had 4,500 customers, said John Montgomery, Microsoft’s corporate vice president.
Most tech companies are currently introducing four types of enterprise generative AI products: features and services that generate code for software developers, create new content such as sales emails and product descriptions for marketing teams, and drill down on company data to enable employees to answer questions and summarize meeting notes and longer documents.
“It’s going to be a tool that people can use to accomplish what they’re already doing,” said Bern Elliot, vice president and analyst at IT research and advisory firm Gartner.
But the use of generative AI in the workplace carries risks. Chatbots can generate inaccuracies and misinformation, provide inappropriate responses, and leak data. AI remains largely unregulated.
In response to these issues, technology companies have taken a number of steps. To prevent data leakage and increase security, some have developed generative AI products so that they do not keep their customers’ data.
When Salesforce launched AI Cloud last month, a service with nine generative AI-powered products for businesses, the company built in a “trust layer” to mask sensitive company information to prevent leaks and promised that user input into these products would not be used would to retrain the underlying AI model.
Similarly, Oracle said customer data is kept in a secure environment while training its AI model, adding that the company cannot see the information.
Salesforce offers AI Cloud starting at $360,000 per year, with costs increasing based on usage. Microsoft charges for the Azure OpenAI Service based on the version of OpenAI technology that a customer chooses and based on the scope of use.
For now, generative AI is being used primarily in low-risk, human-centered workplace scenarios — rather than highly regulated industries — said Beena Ammanath, executive director of the Deloitte AI Institute, a research center at the consultancy. A recent Gartner survey of 43 companies found that over half of the respondents have no internal policies on generative AI
“It’s not just about being able to use these new tools efficiently, but also about preparing your workforce for the new types of work that may be developing,” said Ms. Ammanath. “New skills are needed.”
Panasonic Connect, part of Japanese electronics company Panasonic, began using Microsoft’s Azure OpenAI service to build its own chatbot in February. Today, its employees ask the chatbot 5,000 questions a day on everything from composing emails to writing code.
While Panasonic Connect expected its engineers to be the primary users of the chatbot, other departments – such as legal, accounting, and quality assurance – also turned to it to help summarize legal documents, brainstorm solutions to improve product quality and other tasks, said Judah Reynolds, director of marketing and communications for Panasonic Connect.
“Everyone started using it in ways that we didn’t anticipate ourselves,” he said. “So people really take advantage of it.”