Salesforce TDX reveals how consumer and enterprise AI differ

Salesforce's plans for GenAI endeavors

In a press statement, Salesforce said AI assistants increase the potential for what businesses can do.

"AI assistants can already answer questions, generate content, and dynamically automate actions. And someday, these assistants will become digital sales and service agents, anticipating our needs and operating on our behalf," the statement said.

However, the supplier also said new AI advancements create ethical concerns. "It's one thing if an AI assistant offers a bad product recommendation, but if it takes misguided actions on real-world concerns like personal finances or medical information -- the stakes suddenly become much higher," the statement said.

Salesforce maintains that the new generation of GenAI means it is not realistic to engage in every AI interaction or review every AI-generated output. So, the vendor plans to design more controls so humans are at the helm of AI outcomes and they can focus on tasks that most need their attention.

"In other words, humans aren't always rowing the boat -- but we're very much steering the ship," Salesforce's statement said.

The supplier adduced three ways in which it plans to keep humans at the helm of Salesforce AI: Prompt Builder, Einstein Trust Layer -- which offers an audit trail that enables users to see where their AI assistants have either gone wrong or right -- and Data Cloud.

Overall, the march toward more autonomous AI is clear beneath the metaphor of the human at the helm.

"I can't imagine a single technology when we just say, 'Let it go,' without having some sort of control panel, some sort of oversight, some sort of accuracy around it," said Paula Goldman, chief ethical and humane use officer at Salesforce, in a pre-conference panel for media.

Enterprise GenAI is a different beast than chatbot GenAI

TDX also emphasized that enterprise use cases for GenAI are distinct from consumer use of tools such as OpenAI's ChatGPT and Google's Gemini, formerly Bard.

"The consumer space and enterprise space are very different. In the consumer space, we need to think about an assistant or an entity that does everything. The scenario is very different in the enterprise, where we talk about specific use cases, in a specific domain with specific customers," said Silvio Savarese, executive vice president and chief scientist at Salesforce, in a pre-conference panel.

From a product point of view, the supplier trumpeted Einstein 1 Studio, a clutch of no-code tools that let administrators and developers build GenAI tools for workflows across the platform.

In another panel for press at the conference, Clara Shih, CEO of Salesforce AI, said ChatGPT and similar tools can't answer business-specific questions due to a lack of data and metadata. She added that other copilot tools, such as Microsoft's, do not have the CRM focus that Salesforce's has, as they are more targeted to consumers.

"What we're doing is very different. We don't have an office productivity copilot. Our copilot is laser focused on CRM. If you go and ask any of these other copilots, … 'What's my sales forecast for the quarter?' they won't understand the question because they do not have the data, or the metadata or the business logic," Shih said.

In a press statement, she claimed Einstein 1 Studio would simplify how admins and developers can build and customize Einstein Copilot, as well as embed and customize AI apps in their Salesforce workflows.

Salesforce customers share GenAI experiences

TechTarget's Enterprise Strategy Group published research in 2023 focused on enterprise use cases for GenAI. Overall, the 670 technology and business professionals surveyed globally believed GenAI would benefit customer service the most, followed by marketing, software development, IT operations and product development. More than half of the organizations surveyed expected GenAI to improve process and workflow automation, data analytics and business intelligence initiatives, and employee productivity.

At TDX, Salesforce also marshalled some customer perspectives to back up what its GenAI products can offer.

"We've been eager to tap into the power of generative AI in our business, but data privacy and security concerns were a blocker until now with Prompt Builder," said Andrew Russo, Salesforce architect at BACA Systems. "It allowed us to safely build generative AI experiences that save our users precious time, while still keeping our data secure and inside of our org."

Marissa Scalercio, vice president of sales operations at Carnegie Learning, said: "Prompt Builder allows our team … to bring our CRM data into our prompts easily with no code. This helps us augment our business processes to cut down on service rep response, email outreach and content generation time."

Salesforce does seem to be in a good position to make progress with GenAI for CRM. This, in part, comes down to how it has blended its major acquisitions, such as MuleSoft, Tableau and Slack, into its core CRM business.

I interviewed Tableau's CEO Ryan Aytay at the event, and I shall reflect on that conversation in another blog post.

Brian McKenna is a senior analyst at TechTarget's Enterprise Strategy Group, who focuses on customer experience and other business applications. Previously, he was an editor at ComputerWeekly.

Enterprise Strategy Group is a division of TechTarget. Its analysts have business relationships with vendors.