Salesforce CEO Marc Benioff, one of the tech industry's biggest hype guys, wants you to know that he's passionate about generative AI, but even he doesn't know that generative AI's biggest proponents are I don't think I can accomplish what I say. And he (not surprisingly) blames Microsoft.
A month after Salesforce's huge tech conference Dreamforce, Benioff endlessly touted, touted and praised AI (at least the AI used in Salesforce's own products), it's now time to set expectations. I'm on tour. And he's badmouthing his biggest competitor and biggest rival, Microsoft. He recently appeared on the podcast “Rapid Response,” hosted by Bob Safian, former editor-in-chief of Fast Company.
Regarding the potential of AI, he said, “I've probably never been more excited at Salesforce in my career.”
But he also warned that “customers are being told things about enterprise AI, and perhaps AI in general, that are not true.” “I think Microsoft has done a huge disservice not only to the industry as a whole, but to all the AI research that has ever been done.”
Benioff specifically had negative things to say about the accuracy and usefulness of Microsoft Copilot. He even compared Copilot to Clippy, Microsoft's widely criticized talking clip cartoon in the 1990s that was supposed to be an assistant for Microsoft Office users.
“We are learning from AI priests and priestesses at LLM model companies and Microsoft that AI is curing cancer, AI is curing climate change, and that data centers require us all to connect to nuclear power plants. You may have heard that you have to… None of this is true,” Benioff said.
He was doubling down on the comments he made regarding X. “LLM (Large Language Model) is not a direct bridge to AGI, and much of AI's current potential is simply overestimated. As experts argue, AI is not yet a direct bridge to AGI. We’re not solving climate change.”
This was said by OpenAI CEO Sam Altman. Earlier this summer, he hypothesized that using AI-enhanced health technology:[m]Perhaps a future version will help find a cure for cancer,'' Altman said at the Aspen Ideas Festival, Newsweek reported.
Benioff also cited Gartner's research on Microsoft Copilot. Gartner's report, “Top 10 'pain points' with Copilot for Microsoft 365,” released in April, found that only 4 organizations with Copilot pilot programs currently plan for large-scale deployments. It turns out that only 1 in. This is actually a pretty good number considering how young this technology is and how slow companies have been to adopt it. That said, Gartner concludes that as Copilot improves, so will its adoption in enterprises.
Microsoft may point to market research conducted by Forrester that shows significant benefits for small and medium-sized businesses using Copilot. Based on a study of 266 small and medium-sized businesses, Forrester found that Copilot led to a modest increase in revenue while reducing operating costs and speeding new employee onboarding. However, it should be noted that this research was paid for by Microsoft. Make it what you want.
Still, Benioff points out that today's GenAI, while sometimes amazing, is mostly not ready to replace human workers. Indeed, Google's NotebookLM's podcast creation feature, which lets you create a tongue-in-cheek AI-generated host to explain your content, is quite the party trick. But it's hard to see how such a thing would directly eliminate the soul-sucking menial jobs that stall most corporate jobs.
Benioff is also right that one area where GenAI appears to be doing a very good job for businesses is AI agents. This ties in nicely with Agentforce, a Salesforce product he's been promoting lately. Many other technology companies and startups are also working on AI agent technology, from building use-case specifics to providing platforms that companies can build on their own. Just a few examples include OutRival, Atlassian's Rovo AI, and Sierra, a startup founded by former Salesforce co-CEO Bret Taylor and Google veteran Clay Bavor.
AI agents are especially being leveraged in customer service, from vastly improved website chat bots to field service guides. Examples include Zingtree, Talla, and Neuron7. Salesforce also has products here.
AI agents are also finding success in sales, especially lead generation, the lowest rung of any sales organization: high-rejection cold calls and emails. Examples include AI Regie.ai, AiSDR, Artisan, and 11x.ai.
“I think we'll have over a billion agents using Salesforce within the next 12 months,” Benioff said, based on having about 10,000 customers try Salesforce at technology conferences. .
However, it's also worth pointing out that there are other areas where LLM is already valued and where Microsoft in particular can compete. Software programmers and engineers increasingly use them to aid in testing, debugging, and generating code samples, using tools such as Microsoft's GitHub Copilot. Countless startups also offer AI coding assistants, including JetBrains and Continue, just to name a few.
Microsoft's close relationship with OpenAI means its cloud, Azure, is also a popular choice for companies building their own GenAI apps using the LLM model. So Microsoft's tentacles with GenAI go far beyond creating documents in Word, charts in Excel, and transcribing meetings in Teams.
At the same time, it's safe to say that AI is overhyped when the very people who literally promote it to sell their own AI products are warning that AI is overhyped. there is no.
“Managing expectations while leveraging the power of AI is key,” Benioff explains.