If you had any doubts about just how quickly generative A.I. is going to encroach into our lives, now’s the time to drop them, because Salesforce and Microsoft are injecting the technology into their customer relationship management services—and more.
Microsoft went first yesterday with the announcement of Dynamics 365 Copilot; then Salesforce this morning with the evolution of its Einstein AI into “Einstein GPT.” Get ready to start receiving sales emails that aren’t just A.I.-optimized to reach you at the right time and with targeted content, but that have also been artificially authored. Customer service bots are getting souped up with generative A.I., too.
Microsoft’s Copilot will help you generate product listings for e-commerce sites, and is even being integrated into the firm’s supply-chain management software to “proactively flag external issues such as weather, financials and geography that may impact key supply chain processes”—it can then draft emails to affected supply chain partners, to warn them of disruptions.
Salesforce, meanwhile, is also offering developers an A.I. copilot (though it doesn’t use that term) that can generate code for them. And yes—ChatGPT is also coming to Salesforce-owned Slack.
The big winner here is OpenAI, which is heavily backed by Microsoft, so no surprise to see its technology used on that side. Salesforce’s big pitch is that Einstein GPT will combine Salesforce’s own A.I. models (and of course its own cloud) with external models, such as OpenAI’s—it’s all about the ecosystem, and this is just the first integration.
OpenAI CEO Sam Altman said the use of his firm’s technology in CRM services “allows us to learn more about real-world usage, which is critical to the responsible development and deployment of A.I.—a belief that Salesforce shares with us.” To that end, Salesforce is also launching a $250 million venture fund for the development of “responsible, trusted, and generative A.I.”
It will be fascinating to watch how this plays out. On the one hand, we’re now in the territory of serious businesses using generative A.I. for serious things, as opposed to playing around to see how long it takes to get a chatbot to say something offensive. On the other hand, some of these applications involve customers who may have some curveball questions. And it's worth remembering that generative A.I. technology like OpenAI's ChatGPT will occasionally "hallucinate," that is, basically make up fake information.
In theory, Microsoft and Salesforce’s new offerings should be safer to use because they only draw on information from companies’ own websites and internal databases—the customer-facing elements will in that sense be a bit like those Google search boxes in websites. But that won’t necessarily make these A.I.s immune to occasionally emitting bogus information. Companies will find out soon enough how carefully they need to monitor their new copilots.
Want to send thoughts or suggestions to Data Sheet? Drop a line here.
David Meyer
Data Sheet’s daily news section was written and curated by Andrea Guzman.
