{"id":107766,"date":"2026-10-07T13:21:32","date_gmt":"2026-10-07T09:21:32","guid":{"rendered":"https:\/\/techxmedia.com\/en\/?p=107766"},"modified":"2026-10-07T13:21:33","modified_gmt":"2026-10-07T09:21:33","slug":"agentic-ai-must-earn-autonomy-says-manageengine","status":"publish","type":"post","link":"https:\/\/techxmedia.com\/en\/agentic-ai-must-earn-autonomy-says-manageengine\/","title":{"rendered":"Agentic AI Must Earn Autonomy, Says ManageEngine"},"content":{"rendered":"\n

Agentic AI is changing the role of IT teams, moving them from doing the work to directing it. Speaking to TECHx Media<\/a> at AI Everything 2026, Praneeth V., Technical Evangelist at ManageEngine<\/a>, said organisations should earn autonomy for AI agents step by step, starting with low-risk tasks and keeping human approval for high-impact actions. He also outlined the safeguards needed before agents operate in production, the lessons ManageEngine has learned from integrating AI across IT operations, and the barriers to enterprise AI adoption in the UAE and wider Middle East.<\/p>\n\n\n\n

TECHx Media: AI is increasingly moving beyond assisting IT teams to taking action. How is agentic AI changing the role of IT teams, and how much autonomy should organisations give AI systems?<\/strong><\/p>\n\n\n\n

Agentic AI is shifting IT teams from doing the work to directing it. Instead of manually triaging tickets, chasing alerts or troubleshooting failed patches, IT professionals increasingly define the policies, review the outcomes and handle the exceptions that need human judgement. In effect, their role move from operators to orchestrator.<\/p>\n\n\n\n

That does not make IT teams less important. It makes their expertise more valuable, because someone has to decide what looks good like, where the limits sit and when an agent should step back.<\/p>\n\n\n\n

On autonomy, our view is that it should be earned, not granted all at once. Organisations should start with high-volume, low-risk tasks such as L1 service requests or alert triage, prove reliability and then expand step by step. Higher-impact actions, such as changing access rights or restoring production systems should still require human approval, and every agent action should be auditable. That is the model we follow with Zia Agents at ManageEngine.<\/p>\n\n\n\n

TECHx Media: What has ManageEngine learned from integrating AI across different areas of IT management and operations?<\/strong><\/p>\n\n\n\n

The biggest lesson is that AI delivers the most value when it is embedded in the workflow, not bolted on. We have applied Zia Agents across our USM, FSO and UEMS suites, in every area the gains came from AI acting where the work already happens.<\/p>\n\n\n\n

Second, context matters more than raw model power. An agent investigating an outage or a security alert is only as good as the signals it can correlate across applications, infrastructure, endpoints and identities.<\/p>\n\n\n\n

Finally, trust drives adoption. Customers adopted faster once they could see clear guardrails, full audit trails and a clear commitment that their data is never used to train our models. For many organisations in the UAE, strong digital sovereignty through our Dubai and Abu Dhabi data centres has been just as important.<\/p>\n\n\n\n

TECHx Media: As AI agents gain the ability to access systems and take action, what safeguards should organisations have in place before allowing them to operate in production environments?<\/strong><\/p>\n\n\n\n

Before an AI agent is allowed to operate in production, organisations should treat it like a new privileged user, and arguably with more caution.<\/p>\n\n\n\n