At EY Alumni Week, Experts Seek Responsible AI Adoption

Emmanuel Addeh in Abuja

Business leaders and technology experts have urged organisations to embrace agentic artificial intelligence (AI) responsibly, warning that the shift from AI tools that assist employees to systems capable of executing tasks and making decisions requires clear governance, defined limits and sustained human oversight.

The experts spoke at a virtual webinar organised by the EY Nigeria Alumni Association as part of the EY Global Alumni Week. The session, themed: “Leading with Confidence in the Age of Agentic AI,” examined how increasingly autonomous AI systems could reshape business operations, decision-making, customer service and the workforce.

The webinar featured Wilfred Mamah, Partner and Digital Transformation Services Leader at EY West Africa; Tomiwa Adefokun, Senior Manager, Technology Consulting at EY; and Ashish Bakhshi, Partner and Head of Markets at EY West Africa. The session was moderated by Bisi Onasanya, former Group Managing Director of FirstBank.

Mamah said agentic AI represented a fundamental shift from conventional software that primarily supports users to systems capable of pursuing defined goals, planning tasks, using digital tools and executing approved actions.

According to him, the defining change was not simply improved content generation but the transfer of specific tasks, decisions and actions to software.

“The defining shift is not simply better content generation; it is the delegation of tasks, decisions, and actions to software,” he said, adding that organisations should therefore view agentic AI as a transfer of clearly defined decision rights rather than merely an enhanced chatbot.

He identified banking, insurance and public services as sectors where agentic AI could have immediate applications, including credit and service orchestration, claims triage and settlement support, as well as citizen-service fulfilment.

Mamah explained that AI agents could gather evidence, coordinate approvals, identify exceptions and communicate the next steps, potentially changing not only how individual tasks are performed but also how organisations structure work.

He said the wider impact of agentic AI could be organisational, with work moving away from departmental silos towards business outcomes, routine execution increasingly giving way to exception management and decision-making becoming more continuous and responsive.

Adefokun, however, stressed that the degree of autonomy granted to AI agents should be determined by the level of risk and potential impact associated with their actions.

He said organisations should consider factors including risk, reversibility, confidence and business impact when determining whether an AI system should merely recommend an action, prepare it for human approval, execute it independently while notifying a human, or escalate the matter for intervention.

He urged organisations to establish clear boundaries around the outcomes an AI agent is expected to deliver, the decisions it can make independently, when it must escalate an issue, who is accountable for its actions and the data and systems it is authorised to access.

“The future is not autonomous AI operating without people,” Adefokun said. “It is responsible human-agent collaboration, where technology provides speed and scale while people provide purpose, judgement, and accountability.”

Bakhshi challenged business leaders to move beyond discussions about the potential of AI and translate them into practical changes within their organisations.

He said successful adoption should not be measured by the number of AI agents deployed, but by whether organisations were able to redesign valuable work responsibly while maintaining the confidence of employees, customers and other stakeholders.

The experts said responsible deployment would require organisations to designate accountable business owners for AI agents, establish clear authority and data-access limits, maintain human oversight over sensitive decisions and actions, and ensure that activities undertaken by AI systems could be traced and audited.

They also called for continuous monitoring and formal procedures for approving, updating, suspending and ultimately retiring AI agents as circumstances change.

The speakers warned that poorly governed AI agents could expose organisations to risks including excessive system permissions, manipulated instructions, unauthorised actions, leakage of confidential information, compromised third-party integrations, inadequate audit trails and uncontrolled proliferation of AI systems.

They therefore stressed that as organisations move towards greater AI autonomy, governance and accountability must evolve alongside the technology to ensure that the speed and scale offered by AI do not come at the expense of human judgement, security and organisational responsibility.