7 ways to ensure effective digital leadership in the age of agentic AI

Research suggests pioneering CIOs must blend AI agents with human skills. So how can digital leaders establish an effective balance between agentic and professional capabilities, and successfully lead their people into the agentic-enabled future? Here are seven best-practice tips from the experts.

1. Get past the fear. Huy Dao, director of data and ML platform at Booking.com, says the travel specialist’s data stack makes it easier to develop new AI-enabled services, including a partner-to-guest agentic system for automated communication between customers and hotels.

The firm’s stack includes AWS cloud technology, the Snowflake data platform, ThoughtSpot analytics, and Astronomer orchestration. Dao’s data team also uses the MCP pioneered by Anthropic, the open-source agentic framework LangGraph, and a flexible approach to LLMs.

Dao now has a strong data platform for AI exploration, and he encourages other digital leaders to get involved. “As a business, we must think about how to prepare and ensure our data is ready for agents,” he says. “If your organization is still skeptical about AI, I’d say they should get past that fear because the potential is real.”

2. Focus on amplification. Ankur Anand, CIO at recruitment specialist Harvey Nash, says effective digital leaders will treat AI agents as a new layer of cognitive infrastructure that amplifies human expertise. He says research from thinkers like Anil Seth shows intelligence relies on prediction, interpretation, and context. AI agents excel at the first two, but still depend on human judgment to create meaning. Digital leaders should design operating models in which humans set direction, and agents execute at scale, speed, and precision.

“CIOs can lead their people into an agentic‑enabled future by setting a clear purpose, building trust through governance, and redesigning work around hybrid human‑agent workflows,” Anand says. “They should define decision rights, approved data sources, and guardrails so agents operate transparently and safely, while people retain accountability and oversight.”

3. Make people feel empowered. Ewa Zborowska, research director at tech analyst IDC, says agents are great at automating a discrete task, such as tidying up email replies and allowing professionals to focus on more creative tasks. But things get more challenging when firms join agents together to complete complex, multi-step tasks.

“A lot of organizations see an opportunity to introduce agents, but they also realize it’s not going to be something they can do overnight,” she says. “CIOs will often need agents to work across the full business process, which means understanding how everything works. They’ll need to map the process and ensure they introduce agents in all the right places.”

Zborowska says effective mapping requires careful thought about the people who run those processes, so professionals today aren’t left feeling irrelevant tomorrow. “There’s a huge question about how you manage agents successfully,” she says. “You need to make staff feel empowered with this new technology you’re introducing.”

4. Guide staff maturely. Richard Corbridge, CIO at property specialist Segro, says his organization builds agent capability on a maturity scale. First, they consider AI-enabled processes where the human stays in the loop, and then consider the risks of removing them.

Corbridge says this nuanced approach is crucial because agentic features are included in almost every tech product. He refers to Microsoft Copilot, saying most of Segro’s staff use the tool at least 10 times a day to help them operate more productively.

The general direction of travel, he says, should be to guide people and show how agents can help professionals create additional value. “One of our directors said writing with Copilot is like having a super-intelligent colleague next to you the entire time you’re working,” he says.

5. Work with external experts. Joel Hron, CTO at global content and technology specialist Thomson Reuters, has helped his organization launch Trust in AI Alliance, a forum of senior AI researchers from Anthropic, AWS, Google Cloud, OpenAI, and Thomson Reuters, to discuss building confidence in agentic AI.

Having already led Thomson Reuters’ development of the AI-powered legal research tool Westlaw Advantage, and the firm’s Deep Research agent that reviews insights and strategizes like a researcher, Hron is eager to do more, and that’s where the Alliance comes in.

“Their goal is to exceed human-level performance in all areas of their models,” he says. “Given that we have some of the hardest human judgment-oriented challenges to solve, our work is a good litmus test for how good the models are, and we’ve actually supported many of these companies in terms of evaluating their models over the last three years.”

6. Have tough conversations. Nick Pearson, European CIO at technology specialist Ricoh Europe, says digital leaders must watch how technology companies push agentic solutions.

“Any CIO will tell you they’ve got vendors that are desperate to partner with them, which is great, but they’re also desperate to tie clients in,” he says. “As digital leaders, we’ve got to be careful about how we manage these relationships, and we’re all reflecting on long-term agreements that tie us into a current way of working.”

Pearson’s explorations show that agentic AI can create value. Great digital leaders will steer the business toward these positive results, without being waylaid by FOMO. “We need to look at those tools and see if it’s just a nice thing to have, a hygiene factor, or really adding value — and we’re still having those debates,” he says.

7. Don’t get left behind. Thierry Martin, head of enterprise data and analytics at Toyota Motor Europe, says CIOs must take an engaged approach to agents. While generative technologies like ChatGPT and Copilot can help shortcut laborious tasks, he believes agents offer a bigger boon.

Toyota Europe employees use the Snowflake Intelligence agent to find and exploit enterprise knowledge using natural language. LOB employees can use the technology to generate insights, such as finding regional demands for vehicle specifications, much more rapidly.

Martin is impressed with the results and says other digital leaders should start testing agents as soon as possible. “The priority is to understand AI, and try to get ahead,” he says. “I think that’s the key to success because the pace of change just gets faster.”