In the pre-AI organization, the talent, people, and organization strategy aligned with the work to be done. Managers hired for specific functions and the required skills. Innovation and driving change were their own functions, optimized to bring new products and capabilities to market.   

Leaders need a new talent strategy in the AI era. Many of the tasks people did are now being done by AI agents. Some may be automated and agentic, while most today are human-in-the-middle or augmenting humans, especially in mission-critical areas.

But we’re also in a transformational era, as AI capabilities and how organizations use them evolve in months, not years. We can’t lock into workflows and run a traditional change management program to drive adoption. We’re in an era of continuous change.

We’ve had several recent Coffee With Digital Trailblazer episodes discussing how people must enable AI’s value – and then AI’s value amplifies with talent’s collaboration. I also spoke with several experts about the contributions of leaders and talent in driving AI ROI.     

Creative thinkers raise the questions

According to the World Economic Forum’s Future of  Jobs Report, 66% say creative thinking is on the rise.

In the episode on reskilling mid-career leaders, Digital Trailblazer advisors shared how creative thinkers ask the tough questions.

• Ethical judgment – “We as leaders need to focus on those human skills — skills that machines cannot replace… the critical thinking and most of all the ethical judgment. Those will become a competitive advantage,” said Derrick A. Butts, enterprise CXO AI cyber resiliency advisor, CISO, and founder of Continuums Strategy.
• Beyond productivity drivers – “AI is not just about productivity – use it to unleash your own creativity because that’s the skill that will become in high demand,” Joanne Friedman, CEO and co-founder of ReilAI.
• Finding new sources of value – “Recognizing what it is that you do well and figuring out — being creative — how that can be applied to something else,” says Heather May, founder and president of May Executive Search.

Listen to the episode and review the research slide to see four other skills I recommend for mid-career leaders.

Growing expertise across the organization

Subject matter experts, go-to-people, heroes during a crisis, and idea generators – are they AI supporters or detractors? Do they fear sharing their knowledge will lead to their job loss?

Experts have to do a lot more than train AIs. Organizations need to expand the number of experts who can collaborate and challenge AI agents’ recommendations.

“As AI rapidly reshapes the workforce, companies are facing a defining question: are they building organizations full of passive AI users, or empowered experts whose talent is amplified by AI?” says Sarah Edwards, chief product and strategy officer at Kantata. “Retaining and developing highly skilled talent has become one of the most urgent priorities of the AI era, and the future belongs to companies that can use AI to scale their expertise – and in turn unlock deeper human capability, engagement, and innovation.”

Knowledge graphs, semantic layers, and other tools for building the AI agents’ context are one goal. But today’s value from AI agents is largely coming from human-in-the-middle and human-augmented decision flows. Organizations will need more experts – and in many industries, tribal knowledge and experts nearing retirement age are significant risks.   

Validating AI accuracy is an ongoing responsibility

While CIOs and CAIOs are concerned today about bringing more AI experiments into production, there are two greater challenges.

Once in production, is the AI agent delivering measurable business value?

And as data, models, and decision context change, will we be able to discern when AI agents are delivering accurate and valid results?

“What an AI model can’t do is distinguish good from bad or decide what’s worth acting on – that judgment is becoming one of the scarcest skills in the AI era, and it remains uniquely human,” says Kurt Muehmel, head of AI strategy at Dataiku. “The companies seeing the most ROI from AI are those investing in people who can direct AI effectively and remain accountable for its outputs. A strong team with the right orchestration will get more from an average model than a weak team will get from the best one.”

Agenticops, modelops, and observable AI agents are all pieces of the puzzle needed to monitor AI agents in production. Organizations that deliver AI ROI and avoid the risk of AI debt and AI cost debt will recognize the ongoing role of expert review of AI agent performance.

Building the talent pool of AI agent reliability engineers

Site reliability engineers (SREs) diagnose application issues, find root causes, and recommend remediations. I believe organizations will need AI agent reliability engineers (AREs) to perform a similar function.

“Bad agents don’t make bad choices like people do. Agent misbehavior is a design flaw. It’s a data problem, a governance issue, and it has to be addressed as such,” said Joe Puglisi, growth strategist and fractional CIO at 10xnewco, during the episode on managing AI agents.

Accuracy is just one of the issues. Friedman added, “Employees understand context. They have perspective and apply judgment; they’ll push back if there is an obvious error. Agents need that same level of governance spine. Without it, they may show a high degree of confidence in an answer, but it may be for the wrong problem or a different outcome.”

One solution came from Liz Martinez, managing partner at PMO Whisperer. “You can’t assume AI agents can be assigned all these tasks if it doesn’t have the ability, the exposure, or the experience to do it. And if you don’t monitor, that’s a leadership problem, a business problem, not a technology or an agent problem,” said Martinez.

As any SRE will tell you, their job starts during the development process. The ARE has to collaborate with architects, creative thinkers, and experts on the AI agents’ roles, how they are trained, their decision flows, and decision authorities.  

“To go beyond pure AI agent adoption and achieve actual ROI, organizations must treat AI agents as managed components within software development and delivery workflows, not autonomous actors running in the background without supervision,” says David Colwell, VP of AI and machine learning at Tricentis. “That means assigning each agent a defined role, access boundaries, quality criteria, and escalation paths, in addition to testing the validity and alignment of agents’ actions themselves, not just the code they help produce.”

Developing talent as a force multiplier

In digital transformation, I urge leaders to find force multiplier opportunities that deliver on two or more strategic benefits. Examples are innovation and risk reduction, growth that also drives efficiencies.

In the AI era, CIOs and CHROs partner to find opportunities to experiment and develop talent as force multipliers.

“The real value of AI lies in applying it in practical ways that help people move faster, make better calls, and focus more on work that requires judgment and accountability,” says Keith Moore, CEO at AutoScheduler.AI. “For example, in sectors like warehousing and supply chain, agents won’t replace operators or leaders, but amplify them. AI is the force multiplier that helps teams navigate complexity, act on real-time data, and operate with a level of precision that wasn’t possible before.”

Remember, AI is only reshaping business today. Amplifying AI’s value requires Digital Trailblazers to recognize that growth doesn’t come through automation and productivity alone.

Five things to do today:

1. Review three new leadership skills fo the AI era, plus two others that have changed.
2. Discuss these essential questions for CIOs on planning IT Careers in the AI Era
3. Consider why college grads are pissed off about AI when developing a hiring plan
4. Revisit developing entry-level programs to develop next-generation leaders
5. Redesign the agile operating model for the AI era