Salesforce’s head of AI says the company is frustrated by the unreliability of generative AI, such as hallucinations, or when AI spews out incorrect or biased information, an issue that has prevented many companies from widely releasing content that includes it. product.
Salesforce AI CEO Clara Shih told the forum that this poses a broader problem for enterprises. of wealth Brainstorm tech conference Monday in Park City, Utah: “Do I believe artificial intelligence can drive business?” For many, the answer isn’t that simple. “Companies come to us and they want to know how they can actually deploy these solutions in a way that really makes a difference,” she explains.
Shih believes Salesforce is well-prepared to handle this challenge because customers have entrusted their data and business processes to Salesforce for years. “It provides an ideal foundation for artificial intelligence to really inject the context needed for the model to really be able to perform,” she said.
A few months after OpenAI released ChatGPT in November 2022, Shih was appointed as the company’s first head of artificial intelligence. Core trust issues such as privacy. She emphasized that Salesforce is implementing all artificial intelligence internally, which helps the company understand customer concerns. “We like to drink our own martinis,” Shi joked. “It’s been bumpy at times, but I really enjoy the responsibilities we have as client zero.”
Since launching its original Einstein AI product in 2016, Salesforce has actually been developing artificial intelligence to help customers, who mainly use its software for customer service, sales and automated marketing. Benioff’s admiration for Albert Einstein, Salesforce focused on predictive artificial intelligence, which was cutting-edge technology at the time.
In November 2022, a few months after OpenAI launched the popular ChatGPT, Salesforce launched EinsteinGPT, one of the first chatbots for a major company based on a large language model. Salesforce subsequently launched various other AI products, including Einstein Copilot, Copilot Studio, Prompt Builder, RAG, Hybrid Search, and Einstein Trust Layer.
On Monday, Shih promoted a new planned product called Einstein Service Agent, a chatbot designed for use by customer service agents. The chatbot is trained on company data and can easily hand over its work to human customer service agents if needed. Shih said the focus of chatbots may expand beyond customer service in the future.
But, she added, trust is critical for these types of products, and trust itself depends on the specific context in which these tools are used. For example, she said, “Do I trust AI to help me figure out which customers are on this order?” And using Salesforce’s new chatbot, “Do I trust AI to help me answer customer questions?” Apparently, at least for Salesforce, Shih insists the answer is yes.
CEO Daily provides the critical context news leaders in the business world need to understand. Every weekday morning, more than 125,000 readers trust CEO Daily to get insights on the C-suite and beyond. Subscribe now.
