Earlier this year, AT&T’s vice president of data science, Mark Austin, noticed that some of the company’s developers had started using ChatGPT chatbots at work. When a developer got stuck, he asked ChatGPT to explain, fix, and improve the code.
Austin said this seemed like a game-changer. However, since ChatGPT is a publicly available tool, I wondered if it is safe for businesses to use.
So in January, AT&T trialed a Microsoft product called Azure OpenAI Services that allows companies to build their own AI-powered chatbots. AT&T used it to create its own AI assistant, Ask AT&T, to help developers automate the coding process. AT&T customer service representatives have also started using chatbots to help with tasks like summarizing calls.
“Once they see that they can do it, they love it,” Austin said. Forms that once took hours to complete can be completed in just two minutes with Ask AT&T, freeing up employees to focus on more complex tasks and chatbot developers to increase productivity. He said his sexuality improved by 20 to 50 percent.
AT&T is one of many companies looking to power chatbots and harness the power of generative artificial intelligence, a technology that has excited Silicon Valley in recent months. Generative AI is a feature that can generate your own text, photos, and videos upon your prompts, a feature that helps automate tasks such as meeting minutes and reduces paperwork.
To meet this new demand, tech companies are racing to introduce enterprise products that incorporate generative AI. In the past three months, Amazon, Box, and Cisco have announced plans for products powered by generative AI that can write code, analyze documents, and summarize conferences. Salesforce also recently announced a generative AI product used by its sales, marketing, and Slack messaging services, and Oracle announced new AI capabilities for HR teams.
These companies are also stepping up their investment in AI development. In May, Salesforce Ventures, the venture capital arm of Oracle and Salesforce, invested in Cohere, a Toronto startup focused on generative AI for business use. Oracle also resells Cohere technology.
“I think it’s a complete breakthrough in enterprise software,” Box CEO Aaron Levy said of generative AI, calling it “for the first time in history, you can actually start to understand what’s inside it.” We call this an incredibly exciting opportunity.” Data can be stored in ways that were not possible before. “
Many of these tech companies are chasing Microsoft, which invested $13 billion in OpenAI, the developer of ChatGPT. In January, Microsoft made the Azure OpenAI Service available to customers, allowing them to access his OpenAI’s technology and build their own version of his ChatGPT. Microsoft corporate vice president John Montgomery said the service had 4,500 customers as of May.
Most tech companies currently have four types of generative AI products for enterprises. That means features and services that generate code for software engineers, create new content such as sales emails and product descriptions for marketing teams, and find company data to answer employees. Ask questions, summarize meeting notes or long documents.
“It’s going to be a tool people use to accomplish what they’re already doing,” said Vern Elliott, vice president and analyst at IT research and consulting firm Gartner.
But using generative AI in the workplace comes with risks. Chatbots can generate inaccurate or misleading information, provide inappropriate responses, or leak data. AI is still largely unregulated.
In response to these issues, technology companies have taken several steps. To prevent data leaks and improve security, some companies are designing generative AI products to not retain corporate data, directing AI models to only answer questions based on data sources. I have.
When Salesforce last month introduced AI Cloud, a service that includes nine generative AI-powered products for enterprises, the company built in a “trust layer” that helps obfuscate sensitive company information so that users can promised not to use the information entered into the product for retraining. Underlying AI model.
Similarly, Oracle said customer data is stored in a secure environment while AI models are being trained, adding that the information cannot be viewed.
Salesforce offers AI Cloud starting at $360,000 per year, with costs rising with usage. Microsoft charges for the Azure OpenAI service based on the version of OpenAI technology selected by the customer and usage.
According to Veena Amanas, executive director of Deloitte’s research center, the Deloitte AI Institute, generative AI is currently primarily used in low-risk workplace scenarios involving humans, rather than in highly regulated industries. It has been. A recent Gartner survey of 43 of her companies found that more than half of respondents do not have internal policies on generative AI.
“This is not only about being able to use these new tools efficiently, but it is also about preparing our employees for the new kinds of jobs that may evolve,” Amanas said. . “New skills will be needed.”
Panasonic Connect, part of Japanese electronics company Panasonic, began building its own chatbots using Microsoft’s Azure OpenAI service in February. Now, the company’s employees ask the chatbot 5,000 questions a day about everything from drafting emails to writing code.
Panasonic Connect expected engineers to be the main users of the chatbot, but other departments such as legal, accounting, and quality assurance are also expected to be involved in summarizing legal documents, and improving product quality. Judah said they are looking to chatbots to help brainstorm solutions and other tasks.Mr. Reynolds, Head of Marketing and Communications, Panasonic Connect
“Everybody started using it in ways we never expected,” he said. “So people are really using it.”