COVER STORY
The AI-powered defence layer
WE AT ZOHO AND MANAGEENGINE RUN OWN DATA CENTRES GLOBALLY AND SERVE OVER 125 MILLION USERS. COMPLEXITY IS PART OF OUR DAILY REALITY.
“ We’ ve trained Zia to watch for not just the obvious threats, but the silent ones,” said Ganesan.“ It’ s a layer of intelligence that grows stronger over time.”
Automation that understands risk
In addition to building unified platforms and AI engines, the organisation also implements what Ganesan refers to as‘ intelligent delegation’. Though automation has always been a fundamental aspect of IT, unlike the traditional, fixed workflows, this new approach enables automation to adapt by learning from an organisational context. This means understanding who requires access, when, why and accessing the different risk factors.
It also helps with the growing evolution of AI-powered threats. These, Ganesan added, are evolving at breakneck pace. This is true in the cases of phishing and identity spoofing.
“ It’ s an arms race,” he said.“ And no, we don’ t believe 100 % prevention is realistic anymore, but what we can do and are doing is ensuring early detection and rapid containment.”
ManageEngine has embedded defensive intelligence across different layers. While the AI-generated phishing attacks have become harder to detect, with nearperfect grammar, cloned identities and contextually accurate payloads, ManageEngine employs a multilayered security approach:
a. Identity: Multi-Factor Authentication, behavioural analytics, biometric validation b. Endpoint: Anti-ransomware systems monitoring anomalous file behaviour c. Infrastructure: Anomaly detection, least-privilege access models, continuous monitoring
What stands out is the real-time responsiveness. If a system behaves erratically, Zia alerts the relevant team as it’ s happening.
“ It’ s about reducing time to detection and automating the first layer of triage,” said Ganesan.
What’ s next
Looking forward, ManageEngine is approaching the AI era in two waves:
If we take the example of identity security, a single employee could have multiple SaaS logins, cloud platform accounts and email credentials. Zia, through its integration with ManageEngine’ s Privileged Access Management( PAM), automatically detects over-provisioning and flags accounts with excessive entitlements.
“ We’ ve also introduced CIEM – Cloud Infrastructure Entitlement Management,” Ganesan noted.“ It monitors cloud accounts for unusual usage, suggests revoking stale privileges and even recommends policy corrections all based on real user patterns.”
Conversational AI layer: The upcoming‘ Ask Zia’ feature will act as an enterprise co-pilot across all ManageEngine products. It will not only answer questions but also understand role-based permissions, act within system boundaries and deliver IT insights conversationally.
Domain-specific AI agents: The team is developing intelligent agents for use cases such as FinOps, which monitor and benchmark cloud spending across AWS, Azure and Google Cloud. These agents will detect budget anomalies, reconcile usage and even suggest spending optimisation strategies – all autonomously.
This risk-aware automation drastically reduces administrative overhead while improving security posture. It’ s no longer about automating clicks – it’ s about automating judgment.
“ We’ ve set a clear internal goal: deliver five new AI features every month,” said Ganesan.“ But we’ re not just building for hype, we’ re building for clarity, continuity and control.” p
16 INTELLIGENTCIO MIDDLE EAST www. intelligentcio. com