Generative AI in the Workplace: What Leaders Need to Tell Their Teams Right Now
Most of the communication leaders are giving their teams about AI is making the adoption problem worse, not better.
The message leaders are delivering about AI tends to be either vague reassurance ('AI will make your job easier') or abstract urgency ('AI is transforming everything') and both land badly. Vague reassurance is not believed because employees can already see that some things are changing. Abstract urgency creates anxiety without direction.
What follows is the communication framework I've developed from years of working with leadership teams at organisations including Walmart, PwC, Morgan Stanley, and Pfizer navigating exactly this.
Why most AI communication is failing
The core problem is that leaders are trying to manage emotion rather than share information. But the reassurance is often not credible. When a leader says 'AI will augment your role, not replace it' in a context where the organisation has already reduced headcount in an adjacent function, employees don't believe the message. They believe what they can see.
Leaders are trying to manage emotion rather than share information. The reassurance is often not credible employees believe what they can see, not what they're told.
What leaders need to stop saying
'AI will augment your role, not replace it'
This is too broad to be useful. In some roles, AI is primarily augmentative. In other roles, AI is substantively replacing tasks that used to require significant human time. The more honest framing: 'Some of what you currently do will be automated. Here is our understanding of which parts, and here is what that means for how your role evolves.'
'We're exploring AI to stay competitive'
This says nothing. A communication that could apply equally to every organisation in your industry gives your employees no useful information about what is actually happening inside yours.
'Nothing will change without talking to you first'
This is a promise most organisations cannot keep. Making commitments that can't be honoured is worse than making no commitment it damages credibility when the commitment has to be walked back.
What leaders need to start saying
Be specific about what you know and what you don't
A useful structure: here is what we know about how AI will affect this function. Here is what we are still figuring out. Here is how we will keep you informed as our understanding develops. Here is what you should do in the meantime. That last part is critical the communication has to end with something actionable.
Name the roles and functions that are changing first
Telling your whole organisation that 'AI is transforming the way we work' is less useful than telling your finance team specifically what AI means for the reconciliation process, the forecasting function, and the reporting cycle. Function-specific conversations build more trust than organisation-wide messages.
Talk about the upskilling plan in specifics, not generalities
The question employees are actually asking is: what does that mean for me, specifically? The organisations getting the best adoption rates are the ones with specific, function-level upskilling plans not company-wide digital literacy programmes.
The harder conversations most leaders are avoiding
The conversation about roles that genuinely are at risk
Some roles will be significantly reduced or eliminated as AI scales. Most organisations know which ones. Most are not telling the people in those roles because the conversation is uncomfortable. This is a mistake. The information vacuum gets filled by rumour, and rumour is almost always more alarming than reality.
The conversation about what you're not going to automate and why
Most AI communication focuses on what is being automated. Equally important is what isn't and the reasoning. Which human judgments, relationships, and capabilities are you specifically choosing to protect? These conversations are rarer than they should be and disproportionately effective at building trust.
The conversation about AI and cybersecurity risk
As Agentic AI systems take on more operational tasks managing communications, accessing internal systems, executing workflows the cybersecurity surface area of an organisation expands significantly. Leaders need to communicate specifically about AI data governance: what employees should and shouldn't input into AI tools, what internal data can be shared with AI systems, and what the escalation path is if an AI tool behaves unexpectedly. This is not a conversation to delegate entirely to the IT department; it requires behavioural change across the workforce.
The automation conversation DevOps and operations leaders need to have
For DevOps teams and operations leaders, Agentic AI is changing the architecture of how systems are built, maintained, and operated. The conversations that technical leaders need to have with their teams:
Which automated workflows are now being handled by AI agents, and what does human oversight of those agents look like?
How are we auditing AI agent actions in our production environment and who is responsible for that audit?
What is our protocol when an AI agent takes an action that produces an unexpected outcome?
How are we managing the security implications of AI agents that have access to production systems, external APIs, and sensitive data?
A framework for AI communication that actually works
Specificity over reassurance. Name the actual changes happening in this function, not just the general trend.
Honest acknowledgment of uncertainty. Tell people what you don't yet know, and when you expect to know it.
A concrete action step. Every AI communication should end with something the employee can do.
A feedback mechanism. The communication should be two-way. Create a channel for employees to share how AI is actually affecting their work.
Consistency over time. One all-hands announcement is not an AI communication strategy.
The Culture in an AI World keynote is specifically designed for organisations navigating this communication and adoption challenge building on the frameworks above for leadership teams, CHROs, and conference audiences. For the workforce transformation angle, the Fearless Government keynote addresses how public sector organisations are handling this differently. And for organisations looking at the full picture of how innovation strategy connects to AI adoption, the Strategy is Energy keynote covers the organisational design dimension in depth. You can book directly here to discuss which keynote fits your event best.
Frequently asked questions
What is the most important thing leaders should communicate to their teams about AI right now?
Specificity. The communication that builds the most trust is not a confident prediction about how AI will affect everyone; it's honest, function-specific information about what is actually changing, what the organisation knows, and what it doesn't. Generic reassurance is not believed. Specific information, even when it includes uncertainty, is.
How should leaders handle the conversation about jobs being replaced by AI?
Directly, and earlier than feels comfortable. The information vacuum gets filled by rumour, which is almost always more alarming than reality. The organisations managing the human dimension of AI adoption best are those that have had honest, early conversations about role risk combined with specific transition support.
What should leaders tell their teams about the security risks of AI tools?
At minimum: what data should and shouldn't be entered into AI tools, what internal data can be shared with AI systems, and what the escalation path is if an AI tool behaves unexpectedly. Agentic AI systems that have access to internal systems, external APIs, and sensitive data significantly expand an organisation's cybersecurity surface area. This is a leadership communication issue, not just an IT policy issue.
How do you build AI adoption in a workforce that is resistant?
The resistance is almost never to AI itself, it's to the uncertainty AI creates. The most effective adoption strategies address that uncertainty directly: specific information about what is changing and what isn't, a genuine upskilling plan with function-level detail, and visible evidence that the organisation is making deliberate choices. Psychological safety matters enormously. People adopt AI tools faster when experimentation is genuinely safe.
What is the difference between generative AI and Agentic AI for the average employee?
Generative AI responds when you ask it something you prompt it, it produces an output, and the loop ends. Agentic AI acts on your behalf without being asked at each step; it can plan a multi-step task, execute it across different systems, and make decisions along the way. For the average employee, this means Agentic AI is not just a tool you use, it's a system that may be performing tasks in your workflow autonomously, which changes what you're responsible for and what you need to monitor.