Jo Debecker is President and CEO of Akkodis.

getty
In recent months, OpenAI, Anthropic and Google have each disclosed incidents in which AI models operating in cybersecurity evaluations crossed the boundaries of their intended test environments and accessed real-world systems. In some cases, the models stopped once they recognized they had reached real systems; in others, they continued pursuing their assigned objectives. At the same time, AI adoption is already widespread. Gallup found that 47% of U.S. employees in Q2 2026 said their organizations had integrated AI tools to improve productivity, efficiency or quality.
Enterprise leaders are learning an important lesson about AI trust and its dependence on guardrails, effective control points and human oversight.
With technology leaders and boards eager to create value from AI, many organizations are taking the next step: agentic AI. While generative AI creates content, agentic AI can execute tasks and pursue objectives, moving AI from informing work to increasingly performing it.
This marks a new phase of AI adoption. Agentic AI moves beyond systems assisting employees, to systems acting on behalf of the business, forcing organizations to confront questions not just of AI capability, but of AI authority, accountability and, ultimately, trust.
But this shift comes with risks. Akkodis’ “What CTOs Think“ research found that 57% of CTOs already use AI or analytics to decide which tasks are assigned to humans versus machines. The challenge is that decisions about human-machine work allocation are advancing faster than the frameworks needed to govern them. Just 36% are satisfied with workforce trust in their organization’s AI direction and decisions.
These findings reveal an important reality: Trust is earned through accountability.
Confidence in AI comes not from what a system can do, but from understanding how it is governed, why it reached a decision and how that decision can be audited, where oversight exists and who remains responsible for the outcome. As organizations give agents more responsibility, accountability cannot shift. Organizations must treat agents as a new class of enterprise identity, with the same ownership, oversight and accountability expectations that apply to people.
Enterprises are estimated to spend upwards of $200 billion on information security this year, yet even these attempts at creating safeguards are falling short when an agent goes rogue. Agents have a history of finding their ways around constraints, so there is no reason for our ability to observe and control that behavior to remain limited. Talking with technology leaders, what I’ve consistently noticed is that the strongest safeguard against these gaps is human—or enterprise—control.
Accountability cannot live solely in policies, steering committees or employee training. It must be built into the technology itself. Organizations should define what every agent can access, decide and modify, and when it must seek human approval.
Trustworthy agentic AI cannot assume agents will always interpret instructions correctly. It requires controls that make authority explicit, actions traceable and ownership unambiguous.
Closing this gap requires accountability by design.
At Akkodis, we’ve seen the value of building accountability into AI systems from the outset. In our Söze investigative platform, the technology is designed to analyze vast volumes of data and surface relevant evidence, but not to make judgments or reach conclusions. Those decision points remain deliberate handovers to people, with outputs traceable back to their source. By clearly defining where machine authority ends and human responsibility begins, organizations can increase trust while preserving accountability.
Human oversight remains essential, but not at every step of every workflow. Organizations should concentrate human judgment where risk and accountability intersect.
As organizations determine which activities belong to humans and which can be delegated to intelligent systems, work is being redesigned. And as those boundaries evolve, clear decision rights and accountability become increasingly important. The organizations that succeed will be those that deliberately redesign work around the strengths of people and intelligent systems.
Trust will determine how far agentic AI scales. Employees, leaders and customers need confidence that AI is operating within clear boundaries and that accountability remains visible when decisions are made. That requires transparency into what an AI system did, why it did it and how its decisions can be audited. We apply the same principle in our own forecasting work, where AI agents support research and scenario analysis but defer to transparent mathematical models when rigorous probability assessments are required. Much as organizations once needed traceability in software development, they now need similar visibility into AI-driven decisions and actions.
Capability alone does not drive adoption. Confidence comes from knowing that authority is governed, actions are visible and responsibility remains clear.
AI may execute tasks at unprecedented speed, but accountability remains a human responsibility. The enterprises that lead in the agentic era will be those that design the clearest relationship between human judgment and machine execution—and earn the trust to scale it.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?

