AI Takes Center Stage: The Message Salesforce Must Deliver at Dreamforce

Salesforce CEO Marc Benioff gestures on stage at Dreamforce 2016 in San Francisco, CA. Photo credit: Ken Yeung
"The AI Economy," a newsletter exploring AI's impact on business, work, society and tech.
This is "The AI Economy," a weekly LinkedIn-first newsletter about AI's influence on business, work, society and tech and written by Ken Yeung. Sign up here.

Welcome to a special edition of The AI Economy.” Over the next few days, I’ll devote this newsletter to Salesforce’s annual customer conference. New issues will feature an analysis of the company’s AI announcements that might reshape how organizations think about how they use the technology.

The Prompt

Thousands are set to descend on San Francisco this week for Dreamforce, where the worlds of AI, business, and customer engagement collide. Known for its energy and spectacle, Dreamforce has become more than just a tech event—it’s a vision-setting moment for the future of enterprise technology. With Salesforce CEO Marc Benioff leading the charge, expect bold announcements, the highlight of which will be the formal introduction of the company’s Agentforce platform this year.

Ahead of all the fanfare, I spoke with two analysts who have been observing Salesforce’s developments for a long time to figure out what the company is up to and the key messages it must convey as it embarks on this self-proclaimed “hard pivot” and embrace of agentic AI. In short, Benioff needs to emphasize customer value, demonstrate practical, real-world applications of AI and autonomous agents, and explain why it’s worth the high price tag.

Disclosure: I will attend Dreamforce as a Salesforce guest, and the company will cover all my expenses. Salesforce in no way dictated the content of this post. These are my words.

What is Agentforce?

Salesforce's Agentforce website as of Sept. 10, 2024. Image credit: Screenshot
Salesforce’s Agentforce website as of Sept. 10, 2024. Image credit: Screenshot

Announced in August after Salesforce’s fiscal second-quarter earnings report, Agentforce is billed as a new platform for helping customer companies use AI to automate specific tasks. It’s the next evolution of Salesforce’s AI ambitions, eight years after commercializing its Einstein machine learning offering. As Fortune reports, Agentforce lets users “quickly build and deploy autonomous AI-powered agents…that run on top of Salesforce’s existing apps for businesses and automate customer service tasks.”

Chatbots were the low-hanging fruit in the customer service world, and advancements in machine learning and new language models gave rise to things like digital twins and AI agents. Proponents argue that these typically are smarter and more autonomous and will free up human agents to address more creative and impactful work. However, securely building autonomous AI and training them on proprietary data is challenging for organizations.

The deployment of Agentforce could eliminate this problem once and for all.

Although Salesforce has been criticized for not doing enough to adapt to the generative AI moment, the company hasn’t been resting on its laurels. Under the leadership of Clara Shih, it has significantly scaled Einstein, including launching its Einstein GPT for CRM, the Einstein 1 platform, Einstein Copilot Studio to help customers customize their AI, and debuted Einstein-powered service chatbots. In addition, Salesforce also launched a dedicated AI Cloud product to bring large language models to the enterprise. This, combined with the company’s Data Cloud, will be a critical part of the new reality that Salesforce will introduce next week.

Expect the knowledge sharing around Agentforce to kick into overdrive over the next week as Salesforce tries to convince users that they now have the power to build AI-powered apps specific to their needs and that a new data revolution is taking place.

Salesforce's logo is displayed on the exterior of the company's Bellevue, Washington, office. Photo credit: Ken Yeung/The AI Economy
Salesforce’s logo is displayed on the exterior of the company’s Bellevue, Washington, office. Photo credit: Ken Yeung/The AI Economy

Show Me the Value

“There is this moment right now in AI where we have to shift from experiments,” Liz Miller, principal analyst with Constellation Research, tells me. She believes customers are done with pondering over AI and want to use it in a way that generates a good Return on Investment. “Let me go and see if it can do something that adds to the growth, the profitability, the margin, the value of this relationship I’m trying to build with my company.”

This point is underscored by multiple studies revealing that company executives are bewildered by what to do with AI. “The undertow of AI is frustration, and I think that’s an untold truth that no one wants to articulate right now because CIOs, CMOs, CXOs—wherever you sit on that scale—go into meetings every day, and they get asked by their CEO, ‘What are we doing with AI?’,” Miller remarks. “Everyone wants to know what we’re doing with AI.”

Valoir’s chief executive and principal analyst, Rebecca Wettemann, agrees: “At the end of the day, these folks have to show how their decisions and the way they architect things and drive adoption of things are delivering more value.” She says Salesforce has been addressing this with the launch of its Customer Success Score and Starter offering. But even with these features and capabilities, Salesforce must arm their “champions with those kind of tools to be able to show and communicate value.”

“The job…is to show everybody how everybody else is transforming their businesses with Salesforce AI, and if you are not, you are behind,” Wettemann states. She predicts Salesforce will share “very real stories, demos, customers, dogs and ponies that will show people are actually getting value with [Salesforce] today.” In fact, Wettemann expects Benioff to emphasize the ability to “get there better, faster, cheaper, and with less risk with Salesforce than with any other platform” at Dreamforce.

What Does Salesforce Need to Say?

The biggest expectation people will have is when they can use the product. Salesforce must make its new products accessible promptly through a public beta or generally available. Miller notes that in the past, when the company made major product announcements, analysts often joked that these releases wouldn’t arrive until “five years from now.” However, she credits Salesforce for its recent timing improvements. Nevertheless, “there has to be a lot that’s going GA. There has to be a lot for people to go and try and feel and…implement for themselves today if they choose to.”

Attendees try out demos of Salesforce products at the expo hall of Dreamforce 2016 in San Francisco, CA. Photo credit: Ken Yeung/The AI Economy
Attendees try out demos of Salesforce products at the expo hall of Dreamforce 2016 in San Francisco, CA. Photo credit: Ken Yeung/The AI Economy

Both Miller and Wettemann say Salesforce should also prioritize providing pricing clarity. These analysts expressed concern over the CRM platform’s premium status, believing it has become too expensive for some. Wettemann wants to hear conversations taking place, whether at Dreamforce or “in the near future” about predictable and flexible pricing for customers.

Salesforce provided a certain number of credits across multiple packages for early AI adopters. According to Wettemann, the problem is that nobody knew how many calls they were going to make, decided whether they needed generative AI or identified use cases that would generate significant volumes. As a result, these developers were left with credits they weren’t sure how to use. To correct this, Wettemann hopes Salesforce will start talking about a “more flexible bank of AI credits. So as you, as a customer, identify through pilots and other things where you can actually get the most value, you don’t find yourself locked into a bucket of licensing for one use case of AI that you can’t move over to [another one] if you find that this is really where you’re going to get the bang for your buck.”

Miller has heard similar complaints, though she clarifies the problem isn’t Salesforce’s alone to bear. “We haven’t quite settled into the norm of what AI costs,” and now we’re saddled with “the baseline of all of our software costs, plus this unknown factor of what AI will cost us. Is it by consumption? By prompt? Is it a blanket fee? Do I pay credits? Is it fungible tokens? There is so much that’s happening in the industry that people are waiting for some of that price volatility to die down, but they’re also really not happy with a lot of the vendor lock-ins, the mandatory seat increases, the mandatory licensing fee increases that are above and beyond what we’re seeing in the AI cost increases.”

And with Salesforce customers using Agentforce to automate most of an agent’s work, Wettemann argues that per-seat licensing is not a valid pricing model. “There are going to be some tough conversations with customers about what that looks like and how they think about a more of a consumption or value-based pricing model,” she says. Those looking at AI to increase an agent’s productivity are not necessarily interested in the things that will generate exponential value—the tasks AI made possible that were once impossible because they were cost-prohibitive. An example here might be automating workflows using AI and handling assignments usually done by L1 and L2 contact centers. “That’s where the value goes off the charts,” Wettemann remarks. “But that’s also the use cases where I’m no longer charging on a per-seat basis. So Salesforce and customers have to [talk] about what that model looks like in an exponential world.”

Salesforce CEO Marc Benioff on stage at Dreamforce 2023. Photo credit: Salesforce
Salesforce CEO Marc Benioff on stage at Dreamforce 2023. Photo credit: Salesforce

Dreamforce attendees aren’t the only ones Salesforce needs to convince. It must also show Wall Street that it is moving in the right direction, especially after grueling battles with activist investors. Recent acquisitions of AI voice agent maker Tenyx and data management provider Own are further evidence that Salesforce is serious about AI, but how will investors respond to a big bet that is Agentforce?

Miller believes it’s vital for Salesforce to demonstrate a “consistency of adults behind the wheel” from a technological, marketing and growth perspective. The company needs to “get back into that motion of ‘we are Ohana and family who are all here to show that you are successful.’” In other words, Salesforce needs to show that It has a reliable team that will be there for its customer companies, focused on their growth and success, especially in this new era of AI-powered business.

Salesforce Isn’t About Fomenting Fear About AI

Dreamforce represents an opportunity to show where the rubber meets the road. But, while Salesforce will shed light on the “capacity to create” autonomous agents, Miller doesn’t worry about the negative impact on the workforce, something AI naysayers often bring up. Critics claim the technology will lead to a reduction in the human workforce. If you have an autonomous AI agent handling customer service tickets, what need is there for the human representative?

HAL 9000 from Stanley Kubrick's "2001: A Space Odyssey." Image credit: YouTube screenshot
HAL 9000 from Stanley Kubrick’s “2001: A Space Odyssey.” Image credit: YouTube screenshot

Miller calls that a misnomer and believes this thinking could harm the chances of developing successful AI use cases: “I actually think that it does more to foment fear, and it encourages people not to see successful experiments.” To make her case, she references the HAL 9000 AI from Stanley Kubrick’s iconic film 2001: A Space Odyssey, saying the American Film Institute voted it the thirteenth most terrifying villain. Characterizations like that have perpetuated the belief that AI is evil. “So now, when we sit here and say ‘AI is going to take your job, AI is going to replace you,’ did you notice that people didn’t bat an eye and everyone just went, ‘it’s totally going to happen’?”

“They foment that fear when, in reality, what people are experiencing is that AI allows them to self-define what value they want from any given experience,” Miller remarks. “So, if an employee wants to take busy work off their place, there are tools that can do that. There are business applications backed by AI that can do that. And what I think we’re going to see from Salesforce is the capacity to launch those types of autonomous agents without having…the cavalcade of 100 data scientists along with the prompt engineers and all the expensive people you can hire from MIT having to be within your organization at the same time. It extends that Salesforce ethos of anyone can learn how to do this.”

At Dreamforce, Miller predicts we’ll see “the capacity for people to harness AI even more across the Salesforce platform, to demystify some of this scariness, and really watching…admins get hold of this and creating not only different businesses but different business processes, all in the name of driving that growth within these relationships.”

Editor's Note: Subscribe to "The AI Economy" today to receive additional analysis from Dreamforce, plus a weekly roundup of AI news you may have missed.

End Output

Thanks for reading. Be sure to subscribe so you don’t miss any future issues of this newsletter.

Did you miss any AI articles this week? Fret not; I’m curating the big stories in my Flipboard Magazine, “The AI Economy.”

Follow my Flipboard Magazine for all the latest AI news I curate for "The AI Economy" newsletter.
Follow my Flipboard Magazine for all the latest AI news I curate for “The AI Economy” newsletter.

Connect with me on LinkedIn and check out my blog to read more insights and thoughts on business and technology. 

Do you have a story you think would be a great fit for “The AI Economy”? Awesome! Shoot me a message – I’m all ears!

Until next time, stay curious!

Subscribe to “The AI Economy”

New issues published on Fridays, exclusively on LinkedIn

Leave a Reply

Discover more from Ken Yeung

Subscribe now to keep reading and get access to the full archive.

Continue reading