When agent AI began to penetrate collaborative workspaces in 2024, many vendors positioned it as a solution to chronic workforce shortages. AI agents will supplement and replace human employees. Research studies have found that throughout 2025, employees will resist the adoption of generative and agentic AI in the workplace. Concerns that AI will make jobs more difficult or eliminate them completely. Those concerns persisted when Salesforce CEO Marc Benioff made headlines in September 2025 when he said Salesforce was cutting its support workforce by 4,000 people. “Reducing the number of heads” required to do the same job.
A Salesforce spokesperson clarified in January 2026 that support headcount refers to positions, not individual employees. “Over the past year, we have evolved nearly every part of our business with AI. Thanks to the benefits and efficiencies of Agentforce, we didn't have to fill support engineer roles even as demand for customer support increased. Instead, we redeployed hundreds of employees to high-growth areas of our business and created entirely new roles like deployment strategist, AI conversation designer, and AI architect. ”
By the end of 2025, enterprise AI adoption rates had leveled off. As reported by The EconomistA Stanford University report documents quarterly declines, with 37% of Americans using generative AI in the workplace in September 2025, down from 46% in June 2025. The Federal Reserve Bank of St. Louis tracks the use of generative AI year-over-year and found that in August 2025, 12.6% of working-age adults were using generative AI every day at work, an increase of only 0.5% from August 2024.
After an industry-wide one-two punch of ambivalent employees and skeptical business customers, Salesforce is pivoting by making the case to its employees that AI skills will benefit them. human Work experience. Today, the company announced its AI Fluency Playbook. According to a press release, it is a “practical guide for companies to prepare their employees to confidently collaborate with AI to empower employees and drive business impact at speed and scale.”
In a preliminary media briefing on Wednesday, Salesforce chief human resources officer Natalie Scardino touted AI fluency as a key element of the modern workforce. “The real differentiator, and what will separate the winners in this next era, is not the tools themselves, but how organizations shape their governance and integrate AI into their core workflows. It’s the difference between deploying a tool and actually changing the way work is done.”
The three progressive core components that make up Salesforce's AI Fluency are:
-
AI Engagement: Defined as employee sentiment towards AI.
-
AI activation: The stage of incorporating AI into daily workflows.
-
AI expertise: This stage combines human skills such as adaptability and problem-solving with AI tools to improve productivity.
The role agent AI plays in the employee experience also changes depending on the employee's responsibilities. Salesforce had identified four distinct AI roles: as a tool, as an assistant, as a contributor, and as a collaborator.
Salesforce claims that an AI-powered workplace is good for companies and the people who work in them. Scardino said, “When used well, and we've seen it used in our own company, AI increases employee ownership. It gives employees more control over how they work.”
At the briefing, two Salesforce customers also spoke about how their employees are adapting to this new AI fluency. Pierre Matuché, Senior Vice President of IT and Digital Transformation at Adecco, and Greg Shoemaker, CEO of r.Potential.
Shoemaker talked about how the company used agent AI to create chief potential officers (CPOs). He said this AI is essentially a combination of multiple agents that can contextualize the current workforce structure of all the AI capabilities evolving in the market, allowing leaders to say, “Okay, this is where I am, this is the landscape, and how can I make better decisions.”
Adecco's Matuchet talked about how agent AI has made his company's recruiters more efficient by taking over processes like screening and appointment scheduling.
The consensus between the two customers was that to get the most out of agent AI, companies needed to use the best organizational approach, not the best technical tools. Matuchet said, “Before you can agentize, you have to standardize your processes. We're even using AI as an excuse to ask teams to increase process standardization, because we're explaining to teams that if processes aren't standardized, they'll fail to implement AI.”
The shoemaker agreed. “Technology is not a strategy. Companies that treat this like a software project are struggling, but companies that are looking at this from an organizational capability perspective are doing it right.”
