AI Is Coming. Have the Right Conversation With Your Employees.

Paul Sala
May 14
8 min read

How to use AI without losing your most important asset: employee talent.

By Paul Sala, Co-Founder, Ibex Ascent | ibexascent.com

Your employees already know AI is coming. They can see the headlines. They can hear the boardroom language: productivity, efficiency, automation, higher-value work. They have heard those words before and they know they do not always end in a better working day.

So when leaders announce an AI programme with too much excitement and not enough honesty, employees do not feel inspired. Some feel curious. A few feel hopeful. Many feel cautious. More than a few feel scared.

That fear is not irrational. AI is already capable of absorbing tasks, compressing roles, changing performance expectations, and reshaping how work gets done. It can write first drafts, summarise meetings, analyse data, respond to customers, generate reports, and spot anomalies. It may not replace most jobs in one clean sweep. But it is already coming for the version of many jobs that exists today.

The real employee question is not "Can AI help the company?" It is "What does this mean for me?"

Until leaders answer that question clearly, AI adoption will sit on fragile ground. Because the success of AI will not depend only on models, platforms, licences, and use cases. It will depend on whether employees trust the organisation enough to help make AI work.

That is the conversation leaders need to have now. Not a hype conversation. Not a reassurance campaign. Not another mandatory training rollout dressed up as transformation. A real conversation about work, value, trust, and talent.

The wrong conversation starts with productivity

Most organisations introduce AI through the language of productivity. That makes sense at one level. AI can help people do things faster, reduce manual effort, and improve throughput.

But when leaders lead with productivity, employees often hear something else. They hear: "We are looking for ways to do the same work with fewer people."

That may not be the stated intention. But employees are not listening in a vacuum. They are interpreting AI through years of previous restructures, outsourcing efforts, and efficiency drives. They know the pattern. A new tool arrives. A productivity case is made. Processes are redesigned. Targets increase. Teams are asked to do more with less. Eventually, the workforce conversation follows.

The problem is not that productivity is irrelevant. The problem is making productivity the primary story. Productivity tells you that capacity has been released. It does not tell you whether business value has been created, whether customers are better served, or whether employees are more capable. And it definitely does not tell employees whether they have a future inside the change.

The better question is not "How many hours can AI save?" It is "What business value can we create with the capacity AI gives us?"

That one shift changes everything about the conversation.

Employees are not resisting AI. They are reading the room.

It is tempting to describe employee concern as resistance. That is convenient, but it is usually wrong. Many employees are already using AI outside formal company programmes, experimenting, learning, trying to understand what it can do. The issue is not fear of the technology itself. The issue is fear of what the organisation might do with it.

Employees are asking questions leaders may not have answered:

Will this make me more valuable, or less necessary?
Will I be trained properly, or simply expected to keep up?
Will AI be used to monitor me?
Will my judgement matter, or will I be expected to follow machine recommendations?
Will the value created by AI be reinvested into better work, or converted into headcount reduction?

These are not cynical questions. They are practical ones. AI changes the psychological contract at work. It changes the relationship between skill, effort, output, value, and security. If leaders treat that as a communications issue, they will miss the point. This is a trust issue.

"AI will change work. Some tasks will reduce. Some roles will evolve. We are not going to pretend otherwise. Our goal is to use AI to improve outcomes and make work better. That means employees will be involved in shaping how AI is used, where human judgement remains essential, and how released capacity is reinvested."

That is not soft. It is credible.

The real risk is not rebellion. It is indifference.

In most organisations, the AI programme will not be defeated by dramatic employee sabotage. It will be defeated by quiet non-adoption.

Employees will attend the training but not change how they work. They will activate the licence but use it only for basic tasks. They will sit in workshops but not share the real risks. They will see poor use cases forming but stay silent. They will comply with the process but withhold the creativity, judgement, and local knowledge that make transformation succeed.

That is especially dangerous in a time of low engagement. AI adoption depends heavily on discretionary effort. Employees need to experiment, test, challenge, and refine. They need to explain where work really breaks, where data is unreliable, where automation could create risk. That knowledge does not live in the tool. It lives in the workforce.

A company can buy the platform, launch the pilot, complete the training, and still fail to create meaningful value because the people closest to the work never truly committed.

The leadership question is not just "How do we get employees to use AI?" It is "How do we make AI worth committing to?"

Every AI programme has a workforce bargain. State it clearly.

The bargain might be: "Help us automate work, and we will reduce headcount." Or: "Help us free up capacity, and we will reinvest that capacity into better work, better service, and new opportunities." Employees will try to work out which bargain is real. The organisation should not leave that to rumour.

Senior leaders need to be explicit about what they are asking employees to accept in exchange for supporting AI. That does not mean promising that no role will ever change. It means explaining the principles that will guide decisions.

"Where AI releases capacity, we will look first at how that capacity can be redeployed into named business outcomes. Where roles change, we will redesign work and reskill people before considering role reduction."

The key principle is redesign before reduction. Before headcount reduction is considered, leaders should answer: What tasks have changed? What capacity has been released? Where could that capacity create more value? Which roles can be redesigned, and what training is available?

This does not remove every hard decision. But it creates a more honest and disciplined approach. It tells employees the organisation is not pretending work will stay the same, but it is also not treating people as the first place to harvest savings.

Do not ask employees to adopt AI. Ask them to help redesign work.

One of the biggest mistakes organisations make is treating employees as end users of AI. That is too small a role.

Employees should be treated as co-designers of better work. The people closest to the work understand what no vendor demo can show. They know where the process is broken, which tasks are repetitive and which require judgement, where handoffs fail, where data cannot be trusted, and where automation could do real damage.

If you want employees to support AI, give them a role worth playing. Do not say "Here is the tool. Please adopt it." Say: "You understand the work. Help us decide where AI can improve it, where it could harm it, and where human judgement must remain."

A simple test for any leadership team: Where have employees actually changed the design of the AI programme? If the answer is nowhere, they have not been involved. They have been informed. That is not enough.

Start with work employees actually want fixed.

Start with the friction employees live with every day: too much admin, manual reporting, poor knowledge search, slow approvals, rework caused by bad data, customer queries that bounce between teams.

The first reaction you want is not "Here comes another tool." It is "Finally, someone is fixing the nonsense."

That is how trust starts to build.

Measure value beyond the productivity scorecard.

If the AI scorecard is dominated by hours saved, FTE equivalents, and output per employee, everyone will understand the real agenda. A stronger value story asks: Are customers getting better service? Are errors reducing? Are managers making better decisions? Are risks detected earlier? Is knowledge more accessible?

The aim is not fewer people. The aim is more business value from the people you already have. That is a much more credible and durable AI strategy.

Managers are the bridge between fear and commitment.

Employees experience AI through their manager. A brilliant executive strategy can still fail if managers cannot explain what is changing, why it matters, and how concerns will be handled.

"I know some of this feels uncertain. We are going to look at the work together, identify where AI can remove friction, where it can improve quality, and where human judgement must stay in control. I want this team involved early, because we know the work better than anyone designing it from the centre."

Managers should not be asked merely to drive adoption. They must help redesign work, protect trust, surface risk, and turn AI from a central programme into something meaningful at team level.

Senior leaders must define what AI will not do.

Trust is not built only by saying what AI can do. It is built equally by saying what AI will not do. Will AI be used to monitor behaviour? Will AI usage influence performance reviews? Can employees challenge AI-generated outputs? Who is accountable when AI is wrong?

These questions should be part of the leadership conversation from the beginning. The organisation needs clear rules before trust is tested.

The conversation to have now: five questions every leader should answer.

Purpose. Why are we actually using AI? The answer needs to be specific. "To stay competitive" is not an answer.

Work. How will work actually change? Which tasks may be automated? Which roles will evolve? Employees need operational detail, not slogans.

Value. What are we creating beyond productivity? If the only measure is time saved, employees will correctly assume the real value is labour reduction.

People. What happens to talent? How will roles be redesigned? How will released capacity be reinvested? This is the part employees are listening for most closely.

Trust. What safeguards exist? What will AI not be used for? Who is accountable when it is wrong? Good intentions are not enough. Trust needs structure.

The companies that get this right will have an advantage.

But the companies that win will not simply be the ones that move fastest with technology. They will be the ones that move most deliberately with people.

Employee talent is not a soft issue. It is the asset that makes AI useful. Employees hold the context, judgement, customer understanding, operational knowledge, and practical experience that AI needs to create value safely.

The goal is informed commitment: employees who know what is changing, why it matters, how they can shape it, what safeguards exist, and how they can stay valuable as work evolves.

AI is moving fast. But trust still moves at human speed.

The question is not whether AI is coming. It is whether leaders are ready to have the right conversation with the people whose work, judgement, and trust will determine whether AI creates value.