The Big AI Disconnect Between Leadership and Teams

The gap between what company leaders say they want to accomplish with AI and what their teams know about using AI is enormous. At SXSW 2025 this week, leaders at three workforce-related companies shared their research to quantify the disconnect during the panel session: AI & The Productivity Paradox: Are We Really Getting Ahead?

At the same time, the three executives defined how teams can begin to capitalize on AI to improve the quality of their products and services, increase overall productivity and creativity, and streamline processes for employees.

The purpose of this post is to put a spotlight on the chasm between executive vision and employee reality at this point in time. Clarity around this is critical to advance AI adoption across any organization.

I shared some of the following data with my colleagues at Matador Network to communicate how we’re all figuring this out together. That provided a bit of a pressure release valve for company culture because a lot of employees are using AI, but at the same time there’s presently a strong sense of, what are we really accomplishing? Team members did appreciate seeing the data, though, to back up their feelings about company-wide AI integration and the work still to be done.

The three panelists at the SXSW 2025 session were:

• 
  
  Annie Dean
  
  Global Head of Team Anywhere, Atlassian (project management software)
• 
  
  Kelly Monahan, Ph.D.
  
  , Managing Director, Upwork (freelancer marketplace)
• 
  
  Lucas Puente
  
  Vice President for Research, Slack (team communications platform)

Dean kicked off the session providing some data from Atlassian’s new State of Teams 2025 report, which surveyed 12,000 knowledge workers and 200 executives. She officially launched the report a few days earlier during another session: State of Teams: The Future of Work Arrived, and No One Was Ready For It.

“What we found from that research is that employees have never had more information but they’ve never been less informed,” said Dean. “Two-thirds of employees know that AI is going to transform how they work, going to improve the quality of their work, the speed of their work. And yet, 96% of executives are saying that their teams don’t yet know how to adopt it. So there’s a pretty big chasm between expectations and reality.”

Puente added that Slack surveyed 5,000 knowledge workers. The results showed that about 30% of them are “AI maximizers” who fully embrace AI and use it daily for work purposes. About 50%, he said, “are watching, some with skepticism because they don’t really know where AI is going, and some with interest, but they’re not willing to jump into the deep end until there are clear guidelines from their employer.”

Then there’s the cohort in-between who are using AI secretly.

“What’s interesting is that 20% are using AI but under the radar,” Puente said. “They’re scared to tell their manager or employer, ‘Hey, I’m using AI to get my work done’ because there aren’t clear guidelines in the business around what to do and what not to do with AI.”

Monahan then shared that freelancers on average use AI about two times more than full-time employees. Why is the freelancer market so far ahead of full-time employees?

 “A couple of reasons: Number one, they don’t have the bureaucracy that many organizations have,” she explained. “Lucas talked about how many full-time employees are afraid to use AI or using it under the radar because they don’t have clear corporate policies. The second reason freelancers gave us is that their livelihood depends on this. In order to stay current, they have to make sure they’re providing new value. AI is one of the hottest in-demand skills on our platform.”

Monahan identified how companies have always valued employees who are creative and innovative because they’re the wellspring of future leadership. But with AI, many companies seem to be much more focused on productivity and efficiency gains, presumably because those results are easier to see and measure within today’s highly fluid AI landscape.

Except, trillion-dollar companies like Microsoft, Google and Meta aren’t investing tens of billions of dollars in AI, and pivoting their entire business models around AI, so we can write emails faster.

“We’re at this profound tension and friction point,” Monahan asserted. “If we’re going to get AI right, it has to be so much more than an efficiency play. We have to begin to understand how we unlock human intelligence in our workforce.”

All of the panelists emphasized the massive waste of time we spend looking for information, which crushes how employees unlock human intelligence, innovation and creativity. Also, many employees need to call meetings to find the information they need, adding more delay.

“We found in our latest research that 2.4 billion hours are being wasted every year in the Fortune 500 looking for information,” said Dean at Atlassian. “Employees said their number one barrier to moving fast was finding information, and teams were spending a quarter of their week looking for it… [And], 56% of the 12,000 knowledge workers we surveyed said they need to schedule a meeting to get informed.”

AI is supposed to solve for a lot of that.

It gets worse. Monahan shared Upwork data showing that 77% of employees say AI is actually adding to their workload today, not taking away. And, 47% don’t even know how to gain the productivity their leaders are expecting from them with AI.

“To me, this becomes a scaling issue, the technologies are here,” she said. “Our data has shown that the workforce needs to learn how to learn. That is a lost art in our organizations today, and that is actually where I think we’ve got to be upskilling right now.”

Dean, Monahan and Puente provided a glimpse into how AI will help teams optimize their workflows, help employees decrease stress and burnout, and help them get some of their lives back.

The panelists explained how we’re entering an era where AI can capture all company emails, internal content and data, and employee meetings and messages depending on how and where those are all collected. An AI can synthesize that aggregate of information, provide daily summaries, and help identify opportunities and challenges to inform overall strategy.

On a tactical basis, such an AI mothership can help employees find what they’re looking for instantly, versus wasting endless hours searching for things. It will also help reduce duplicative efforts across the entire company. The Atlassian State of Teams report shows how approximately one in two employees agrees that teams “tend to unknowingly work on the same things.”

“Our recent research found that one of the most profound ways to implement AI at companies is to change our relationship with information,” said Dean. “What that takes is remembering that AI runs on digital interactions. It runs on chats, work products, transcripts from your voice calls. In a corporate context, we need all of that information to be available in a digital system. We like to say that AI doesn’t care what’s happening at the water cooler. It cares what’s happening in your chat messages.”

This is not fantasy. Slack AI is an early precursor of this, where the add-on AI functionality brings together the collective knowledge in an organization based on the projects, data and conversations collected in Slack.

“For example, in Slack, if I’m trying to figure out what’s going on with a key business metric, I could go to the data scientist in charge of that metric and ask them, leading to a long back-and-forth to schedule a meeting,” said Puente. “Or I can just ask Slack’s AI search, and it instantly gives me an answer and clarity.”

Likewise, Atlassian’s Rovo AI platform is designed for enterprises to “seamlessly integrate with Jira, Confluence, Loom and Bitbucket, along with essential platforms like Google Drive, Microsoft SharePoint, Figma, Slack, Microsoft Teams and GitHub.”

Providing further context about Rovo, Dean said, “If you have an AI infrastructure, you can get informed instantly and save tons of time, which you can then use to apply human intelligence—be creative, get focused, get in the flow…. That’s profoundly different than today, where if you want to learn about somebody else’s work, you have to schedule a meeting, bring four people together for 30 minutes scheduled for three days from now, and then that gets rescheduled, and suddenly 12 days have passed and you’re still looking for the information.”

A lot of us can relate to that.

Dean continued discussing Rovo, describing it as, “our surface area of AI across all of our tool scape, including third-party tools. I can just say I’m looking for the document that Abby and I were looking at yesterday, and it just pulls it up for me with a summary, and I’m set to go. I don’t have to ask to get on a call. So I think that’s a very transformative use of AI across companies.”

1. Employees must feel educated and empowered to integrate AI processes into their daily workflows. Clarity and transparency around that, supported with data such as what’s provided here, is crucial for success.
2. There is so much leadership focus on using AI to increase productivity. There should be an equal emphasis on creativity, innovation, unlocking human intelligence, and how we “learn to learn” in a new AI landscape.
3. The majority of AI solutions that collect all company content and data in centralized knowledge bases to optimize strategy are designed for enterprise use. The Slack AI tool, however, is something I’m exploring for use in smaller companies.