‘How Do We Even Start?’: The Agentforce Question That Dominated Dreamforce

Within the tech space, and by extension, the AI space, there are so many gaps acting as the proverbial potholes in the road to success. The AI skills gap. The AI ethics gap. And now, the AI adoption gap. 

As customers, builders, and users, we get the memo by now – adapt or fall behind. Accept that the change is here, and that it is here to stay. We get that. But do the companies in pursuit of innovation at breakneck speed – Salesforce included – understand that whilst they’re inching closer to the finishing line, many of us still haven’t made it past the first turn?

The Long Road to AI Relevancy

Ever since Salesforce launched Agentforce in 2024, it’s been facing an uphill battle for AI relevancy. This is not a problem unique to Salesforce either – once every CRM company, AI vendor, and beyond began to understand just how lucrative AI was going to be, they subsequently jumped on the bandwagon. 

With Agentforce, the challenges were very apparent. From confusion over its pricing, rebrands, and the wealth of new features and tools being announced, to concerns that adoption was lagging, Agentforce has become less of a success story and more of an omnipresent thing to accept. 

Even when its hype cycle died down, the battle for relevancy persisted. Agentforce currently has more than 29,000 paid deals, yet its penetration is still low. Salesforce is estimated to have over 150,000 customers. Agentforce’s adoption rate could sit around 20%, but some estimates suggest that only about 5.3% of Salesforce customers have adopted the product. 

And if we look past Agentforce, it’s clear that Salesforce has chosen a new strategy when it comes to getting ahead with AI. In the last year especially, its partnerships with leading AI vendors like OpenAI and Anthropic have worked hard to establish the company’s potential by sitting at the same table as the greats. Claudeforce almost saved Salesforce’s stock reputation overnight. Having Sam Altman, Dario Amodei, and Jensen Huang at Dreamforce got every major tech publication talking about Salesforce. 

The company’s marketing is doing the heavy lifting in terms of sweetening the deal with investors, critics, and analysts. But closer to home, behind the screens of the people actually responsible for deploying and managing any Salesforce AI rollout, the gaps have been getting larger and larger. 

The AI Gaps

Ever since AI began becoming a major part of enterprise work, it was clear that there were areas that the traditional tech landscape needed to focus its training efforts on. Although data within certain Salesforce career roles, like architects, show that the AI skills gap is narrowing, research from Salesforce shows that roughly 60% to 62% of employees report lacking the skills to use generative AI safely and effectively, while around 70% of leaders agree their teams lack the necessary readiness. 

Then there’s the ethics gap. Trust is not only one of Salesforce’s core values – it is also one of the biggest areas of concern for its customers, especially as AI gets intertwined with so many day-to-day workflows through the likes of AIforce, Claude, and Slackforce. 

After a messy security rollout earlier this year, the pressure to make sure that data is protected, used fairly, and that appropriate guardrails are put in place with AI is stronger than ever.  

‘There’s a Big Gap Around Adoption Nobody Wants to Talk About’

Earlier this year, I wrote that I believed Salesforce’s biggest problem when it came to its AI wasn’t slow adoption, but AI exhaustion. As time has gone on, it’s become clear that AI exhaustion was a growing issue within the Salesforce ecosystem and beyond. However, adoption is also a continuous problem, with AI exhaustion acting as a driving force behind it. 

Cecilia Chiderski, a Salesforce Architect and Director of AI at CLD Partners, told SF Ben that whilst attending Dreamforce, she witnessed firsthand how big of a problem this currently is. 

“There’s a big gap around adoption that nobody wants to talk about,” she said. “This really clicked for me at Dreamforce. I was part of a panel where we talked about lessons learned from Agentforce and Data 360 implementations, and every question from the audience was about adoption. How do we get this in front of our people? How do we even start? How do I get my manager to understand the value?

“These are people who understand the technology very, very well. So here we are, in the middle of the agentic city, with all these new AI offerings and demos around us, and people are still saying, ‘I don’t know what to do with this’.”

She explained that on one hand, in the ecosystem, you have one group of people who understand the technology and can build great things. They alone are not enough – they may not understand the full scope of the business. Then you have another group of people who know the business but may have just been told that all their organization’s problems can go away if they become agentic. 

“Nobody is connecting the two,” she said. “That’s where more discussions need to happen. That’s the gap we need to close.”

Where Do We Go From Here?

So, how do we close this gap?

It can be difficult for the C-Suite of Salesforce to understand what the smaller practitioners are feeling down at the bottom of the ladder. Although the company has made substantial efforts to connect with their community over the last year, it’s evident that they’re still not fully seeing the human side of their new technology. 

“My take is that we can’t forget we’re still talking about people,” Cecilia told me. “There’s all this effort to get people to learn the product, learn the new tools, adopt pre-packaged agents. Do we really think that will do it? Everybody knows by now that the technology works and the models perform. We assume people will just get used to it, that they’re good with technology and they’ll jump on the next wave.”

“People hear all these announcements, and not only are they not there yet, but every announcement also reinforces the feeling that they are late to the party. They’re still battling internally with what to do with all this change.”

For professionals and organizations feeling the pressure, Cecilia recommends setting aside that pressure to adopt AI temporarily in favor of focusing on the organization and people within it as a whole.