Salesforce hails 'next wave' in gen AI with custom agents

Salesforce CEO Marc Benioff has long touted the company’s advantage when it comes to enterprise AI, pointing to the sheer amount of data it has access to.

Its new platform Agentforce looks to capitalize on that by deploying generative AI-driven digital agents grounded in companies’ customized data. Agentforce is at the center of Salesforce’s annual Dreamforce Conference, which attracts tens of thousands of its customers and partners to San Francisco every year.

As the world’s largest provider of customer relationship management software, Salesforce (CRM) is betting on Agentforce to drive the next wave of growth, as it looks to keep pace with rivals, including Microsoft (MSFT). Sales and service agents on the platform operate autonomously 24 hours a day, handling basic inquiries that can then be handed off to human agents. Utilizing every piece of information stored in Salesforce’s Data Cloud, generative AI enables the development of custom agents that can respond to conversational language, allowing robots to learn and reason in real-time, unlike chatbots that respond to pre-determined questions.

Clara Shih, CEO of Salesforce AI, said the new agents eliminate the frustration stemming from old bots.

“It would ask us the same questions over and over again,” she said. “These service agents from AgentForce, they really understand a wide range of human language and vernacular and can map that to what are the trusted data sources.”

The latest product launch comes as Salesforce looks to gain momentum following slowing sales growth stemming from a pullback in customer investments. A handful of companies including Disney (DIS), OpenTable, Kaiser Permanente, and ADP (ADP) have piloted Agentforce so far, but Salesforce is targeting wide adoption, after showcasing the technology at Dreamforce. Benioff said he’s aiming to have one thousand agents deployed by the end of the event.

Video transcript

The sales force touting its new autonomous agents as they like to call it the next wave in artificial intelligence.

Those digital assistance and its platform, Agent Force are at the center of the company's annual Dream Force conference this year.

If you go Fujita joins us with an early look, hey, there to you Maria, in many ways, generative A I has already streamlined the way we work, whether it's analyzing our data or writing and summarizing our emails.

But sales force says this Agent Force platform marks another leap from generative A I to what's known as agentic A I.

So we're talking about agents that can be built and deployed and actually reason on their own.

I sat down with sales forces A is Ceo Clara shy as well as product SVP Adam Evans to get a sneak peek at this technology and why they say it's gonna be driving a hybrid workforce where humans and A I agents work side by side.

What an agent is, is when you start to use the A I, not just for a single task, write me a sales email, summarize this call transcript, but actually an agent that can think and that can come up with a plan of action and then take action.

Those autonomous A I agents are taking shape on a new platform called Agent Force.

It allows companies to build digital agents customized to their business, drawing on every piece of information on data cloud.

Hi, I'm Sophie your fa A I assistant.

Those agents then interact directly with customers.

I can come in and say something in natural language like uh um I I haven't received my solar panels.

Yeah, salesforce product SVP.

Adam Evans walks me through a demo pretending to be a customer of a solar panel company.

Now it's asking me about my email address with evans' account verified.

The agent goes to work and what you're going to see the agent do is now go use those actions to look me up through a separate zoom call.

Evans breaks down the agent's reasoning as it retrieves the product information and explains how to set up an EV home charger.

It then helps the customer set up an appointment for installation.

They can use their actions and connect to their data with those guard rules of policy in a way that's dynamic based off of what they're learning as they go and then they can come back to the customer to help solve the problems.

The agents tackle multiple roles.

A service development rep agent communicates with new buyers.

A sales coach agent helps human customer service reps, hone their own sales pitch and coaches them during the actual call.

There's even an agent that helps develop marketing campaigns.

There's no coding necessary, reducing the cost in time for deployment.

It's a much better experience for the customer because the customer can use natural language.

It's not as frustrating the the old bots, you know, we used to be a frustrating experience.

It would ask us the same questions over and over again.

Now these, these service agents from agent force, they really understand a wide range of human language and vernacular and can map that to what are the trusted data sources.

What are the trusted workflows and business logic that it's authorized to invoke based on who is asking and it's able to automate those actions.

So it's very much here today in terms of um what the agents that are already being deployed.

There's a number of companies that are already using this, including Disney AD P, Kaiser Permanente, as well as Open table Work day Saks fifth, you see the names on the screen there.

Um In terms of what this means for enterprise customers, Ceo Mark Benioff says agents have been able to resolve up to 70% of customer inquiries.

You total all of up.

He says it's a savings of hundreds of millions of dollars for these companies.

He said yesterday and talking to reporters that he hopes to have 1000 agents ready to go out the door deployed by the end of Dream Force next week, guys growth there in Akiko.

Great stuff.

Thanks so much for bringing us that.

And of course, we look forward to any developments out of Dream Force next week.