With AI, your workforce is becoming hybrid

The workforce inside many organisations is beginning to change in a fundamental way. Work that once moved exclusively between people can now move between employees, AI systems and increasingly autonomous agents.

For leaders, this creates questions that reach well beyond productivity. The design of work, the responsibilities given to machines and the capabilities people need will all change as human and AI contributions become more closely connected.

In Episode 22 of The AI Values Podcast, Data & AI Leader Tim Carmichael joins hosts Edosa Odaro and Lindley Gooden to explore what leadership looks like as organisations enter this next stage of AI adoption. Drawing on experience that includes the British Army, work in the Middle East and building an airline from the ground up, Carmichael brings the conversation back repeatedly to leadership, communication and the human consequences of technological change.

His description of the emerging workforce is direct:

“I think you can just consider your workforce as being a mixture. It's a hybrid mix now. It's human and non-human workforce.”

That shift requires leaders to think differently about who, or what, performs work and where human capability creates the greatest value.

Leadership still begins with people

Carmichael's approach to leadership was shaped long before the current wave of AI.

He traces it back to his experience at Sandhurst and the principle of "serve to lead". For him, leadership involves putting the demands of the task and the needs of the team ahead of the leader's own interests.

“So this concept of servant leadership, where it's not about you, it's about the needs of your team or the demands of the task and the needs of the individuals within your team, they all matter more than your own needs.”

AI makes that principle more consequential.

Leaders are making decisions that can alter people's roles, change how performance is measured and transfer activities from employees to machines. Those choices may create measurable efficiency, but employees will experience them through a much more personal question: where do I fit?

That means AI leadership requires more than selecting technologies and approving investment. Leaders need to understand how people experience the change and provide a credible view of what human contribution will look like as work evolves.

Carmichael describes the human layer as one of empowerment. Productivity and efficiency can be measured relatively easily, while empowerment depends on whether people gain the ability to contribute in more valuable ways.

Give transactional work to machines

The arrival of AI creates an opportunity to reconsider how work is divided.

Carmichael draws a relatively clear boundary around highly transactional activity:

“If you're working only in a very transactional ones and zeros type space, give it to a machine to do.”

The implication for workforce strategy is significant.

Organisations have historically designed jobs by grouping activities that need to be completed. AI allows leaders to examine those activities more closely and determine which require human judgement, relationships, interpretation or creativity and which can be handled effectively by machines.

That can change the unit of workforce planning. Instead of assuming that an entire role should remain unchanged or disappear, leaders can redesign the combination of activities performed by people and technology.

Done carefully, this can create more space for human contribution. Employees can spend less time on repetitive work and more time applying judgement, solving unfamiliar problems and working with other people.

The quality of that redesign will matter. Removing tasks without considering what remains can leave employees with fragmented roles or diminish the parts of work that gave them satisfaction and purpose.

The leadership responsibility therefore extends to deciding what humans should have more opportunity to do once machines take on more of the transactional work.

Different perspectives create better teams

Carmichael uses a meteorological analogy to describe what can happen when different perspectives meet.

Salespeople may see an AI opportunity through revenue and savings. Technical specialists may see constraints imposed by the science. Their perspectives can create friction because each group is looking at the same opportunity from a different position.

Carmichael sees value in that friction.

“It's the power of the team that becomes greater than the sum of its parts.”

The aim is to bring people with different strengths into the same process. Technical specialists understand how systems work and where their limitations sit. Business-facing colleagues understand customers, organisational priorities and the outcomes people are trying to achieve.

AI adds another perspective to that team.

Carmichael describes organisations bringing together people and machines with genuinely different capabilities. The value comes from designing those combinations so that each contributes what it does best, while recognising that no individual or AI agent possesses the complete answer.

For leaders, this suggests that AI adoption should be considered partly as a team-design problem.

The organisation needs technical knowledge, business understanding and human judgement to meet in ways that produce better decisions. Creating those connections may matter as much as the capability of the AI itself.

Translation determines whether AI gets adopted

Technical capability can remain invisible when the people who understand it cannot explain why it matters.

Carmichael describes specialists who are highly capable within their field but have little interest in selling or promoting their work. That can create a gap between the people building AI and those deciding whether it deserves investment or adoption.

His solution is practical: pair specialists with people who can translate.

“Two is stronger than one in this case.”

Product leads, project leads and internal consultants can help connect technical expertise with the language of the business. They can explain how a system works, what problem it addresses and where it may be useful, while the specialist provides the depth required to make those explanations credible.

This is an organisational capability in its own right.

Boards rarely need every technical detail behind an AI system. They do need enough understanding to make informed decisions about investment, risk and expected value. Employees also need to understand how a new capability relates to their work if they are expected to adopt it.

Translation therefore connects technical possibility with organisational action.

Without it, valuable work can remain trapped inside specialist teams.

Leaders need to show what AI can do

Carmichael also challenges the assumption that good work will automatically attract attention and investment.

Earlier in his career, he says he would have preferred to let the work speak for itself. His position has changed.

“Basically, show, don't tell.”

For AI leaders, this means making useful applications visible.

Carmichael talks about developing what he calls a "shameless gene": becoming more proactive about explaining, showcasing and demonstrating work to different parts of the organisation. Each audience may recognise a different opportunity once it can see the technology working in context.

This matters because adoption is partly social.

People need reasons to believe that a new system will help them. Executives need to see how an investment connects with organisational priorities. Employees need enough understanding to see where a capability fits into their own work.

Demonstration can build that understanding more effectively than abstract claims about AI capability.

It also creates a discipline for AI teams. Showing the technology in use forces the conversation towards utility. The focus moves towards what the system actually enables people to do.

Human judgement becomes more valuable

The possibility of machines supervising or directing human work introduces a more difficult dimension to the hybrid workforce.

AI systems are already capable of outperforming people in some narrowly defined activities. Carmichael discusses medical examples where AI can support diagnosis, allowing specialist clinicians to concentrate more of their attention on prognosis and treatment.

Yet greater machine capability should not encourage people to surrender their judgement.

Carmichael puts the warning plainly:

“Don't throw your brain away when you're using AI.”

AI systems make errors. Their outputs depend on how they have been trained, configured and applied. As organisations give systems greater responsibility, people need enough understanding to recognise when an output deserves challenge.

This becomes especially important if AI begins coordinating work or supervising elements of human activity.

The governance question then concerns authority. Organisations need clarity about which decisions machines can make, which require human involvement and who remains accountable when human and machine contributions combine.

A hybrid workforce still needs identifiable responsibility.

Listen to Episode 22

‍

Going deeper: the hybrid workforce changes accountability

This episode points towards a shift that workforce strategies and AI strategies will increasingly need to address together. Organisations are moving towards environments where work can be performed by employees, conventional software, AI assistants and autonomous agents within the same operating model.

That changes how leaders need to think about roles.

Job descriptions built around collections of human tasks may become less useful when some activities can move between people and machines. Workforce planning will increasingly involve deciding which capabilities should remain human, which can be automated and where combinations of people and AI produce better outcomes.

This creates an accountability challenge alongside an efficiency opportunity.

When an employee performs a task, organisations usually know who holds the responsibility and where that person sits within the management structure. When an AI agent performs part of the same process, responsibility can become less visible. The system may have been selected by one team, configured by another and used by employees elsewhere in the organisation.

Boards will need clarity about those relationships as agentic systems become more common.

The human consequences also deserve attention. Employees need a credible understanding of where they fit within the emerging operating model. Simply telling people that AI will increase productivity provides little guidance about how their own contribution is expected to change.

Leadership therefore needs to connect automation with capability development. When machines absorb transactional activities, organisations should be deliberate about the human capabilities they want to strengthen in response.

Judgement, communication, interpretation, creativity and the ability to work across different disciplines become particularly valuable in the model Carmichael describes.

The same principle applies to AI itself. Machines can contribute capabilities that people cannot easily replicate, but they still operate within systems designed and governed by humans.

A successful hybrid workforce will depend on how deliberately organisations design those relationships.