This year at GITEX GLOBAL, the conversation isn’t about if AI can drive autonomy in the enterprise—it’s about just how close we are to seeing agentic AI act independently and responsibly in high-stakes environments.
For much of the past decade, enterprise AI was synonymous with digital assistants—software agents tasked with helping humans, not replacing them. But as the dust settles on early successes in automation, a new chapter is opening: Agentic AI, a breed of systems empowered to tackle deeply operational tasks and, in some cases, make decisions autonomously.
The Reality of Autonomous AI Agents
“Truly autonomous agents are no longer hypothetical,” says Levent Ergin, Chief Strategist for Climate, Sustainability, and AI at Informatica. “Across our customer base, we’ve helped enterprises evolve from using AI as an assistant to empowering it as a decision-maker. Take the insurance sector, for example. Leading insurers are now deploying generative models, computer vision, and predictive analytics for high-impact use cases such as intercepting fraud at the point of first notice of loss (FNOL), to flagging high-risk cases, and even autonomously resolving straightforward claims.”
That’s not just industry hype. In banking, agentic AI models are now synthesising vast financial records, evaluating risk profiles, and accelerating lending—all with smoother workflows and faster customer turnaround. In logistics, agentic systems spot supply chain bottlenecks before they materialise, automatically re-routing shipments. Healthcare is also seeing progress, with autonomous agents triaging routine queries, reviewing medical imaging, and escalating only complex cases for human evaluation.
How AI is Elevating the Human Loan Officer Experience
Caveats: Why Human Oversight Is Still Crucial
But the promise of fully autonomous systems comes with a caveat. “While these advancements have been both rapid and remarkable, it remains essential for organisations to establish clear processes for human oversight, ensuring that complex, ambiguous, or edge cases are managed without introducing new risks,” warns Ergin.
Recent studies echo this caution. In insurance, while AI can rapidly process well-structured claims, there have been instances where ambiguous cases—such as cross-jurisdictional disputes or policy exceptions—demanded nuanced human review. In banking, agentic AI flagged borderline loan applications for manual audit, successfully catching edge-case fraud the model might have missed. The lesson: operational autonomy must be paired with robust channels for escalation and override.
What Makes Enterprise-Ready Autonomy Possible?
“To ensure that autonomous AI agents behave as intended, you really need both, a strong data foundation and a solid governance framework,” emphasises Antonio Rizzi, Area VP at ServiceNow. “On the data side, ServiceNow can provide that high-quality data through what we call our Workflow Data Fabric, ensuring the AI is working off reliable and well-organised information. Then, on the governance side, you can use the AI Control Tower to make sure that all the AI’s actions align with corporate policies and regulatory requirements.”
This dual blueprint—combining trusted data pipelines with centralised oversight—defines the new standard. Across industries, IT leaders are investing in workflow orchestration tools and AI control towers to log every agentic action, validate alignment with rules, and enable human review for outlier decisions. High-quality, curated training data ensures that agents act on current, relevant facts, while governance frameworks keep their outputs within legal, ethical, and business bounds.
Agentic AI at Work: When Enterprise Assistants Need Autonomy—and Permission
Looking Ahead: The Roadmap and Unanswered Questions
So how close are we to agentic AI with truly operational autonomy? From retail and insurance to logistics and finance, the prototypes are already live. Yet the unanswered questions loom large: How do we guarantee responsible escalation for edge cases? Will regulatory frameworks keep pace with evolving AI capabilities? And how do we prevent a “black box” scenario where decisions lack transparency and auditability?
The future will hinge on continuous dialogue between technology providers, business strategists, and regulators. If enterprises want to harness agentic AI at scale, transparency, oversight, and well-designed systems must remain centre stage.
As agentic AI steps from the lab into the boardroom, the true measure of progress will not be the number of decisions it can automate, but the trust and control it inspires among the humans it was built to empower. GITEX GLOBAL has set the stage: the shift from co-pilot to decision-maker is well underway—and the next chapter is being written by those who build with caution, vision, and rigour.
