AI may be the biggest wildcard in enterprise technology today, but for ManageEngine CEO Rajesh Ganesan, one aspect is already clear: AI agents must be governed with the same rigour as users, devices, and applications.
In an interview with Enterprise IT News, Rajesh outlined a conceptual but concrete direction – an “agentic control plane” – or what he also calls an “agent ops management system” – designed to help enterprises manage AI agents across their digital infrastructure.
AI agents as a new kind of identity
From ManageEngine’s perspective, AI agents are not mystical black boxes. They are, as Rajesh puts it, “a new entity” within the IT environment.
“At the end of the day, an AI agent is nothing but a piece of software,” he said. “It’s a piece of algorithm executing some automation thrown in with a lot of intelligence, because we have the models giving all the intelligence behind.”
Just as organisations today manage human identities, devices, applications, and network components, Rajesh believes they will soon need to manage AI agents with the same discipline.
That means giving agents their own identities, clearly defined entitlements, continuous monitoring and management, strong governance around what they are allowed to do.
Avoiding “rogue” AI behaviour
Rajesh frames the core questions that an agentic control plane must answer:
- Why does the agent exist?
- What exactly is it allowed to do?
- Is it staying within those bounds, or going rogue?
“When AI agents come into the digital infrastructure, they need proper entitlements defined,” he explained. “The agents need good monitoring, good management, good governance… to monitor if the agent does only what it is allowed to.”
This reflects a broader concern he has about the unintended consequences of rapid AI adoption – not only operational risks, but also business and strategic risks.
You have to accept what you know and also accept what you do not know. Try to find the middle ground.Rajesh Ganesan on AI
He pointed to examples of companies that rapidly embraced AI, laid off staff in anticipation of AI-powered efficiencies, and then were forced to rethink once they confronted the realities of model usage and token costs. Add to that geopolitical questions about who owns and controls AI technologies, and it is clear that enterprise AI is still very much a work in progress.
Building confidence for CIOs and CISOs
For ManageEngine, the goal of an agentic control plane is straightforward: give CIOs and CISOs enough visibility and control that they can confidently allow AI agents into production environments.
“The roadmap for ManageEngine is when something like an AI agent comes into the digital infrastructure, we will give it the equal amount of focus and attention it requires for running your IT operations and security operations,” Rajesh said.
In other words, the company wants to extend its existing strengths in IT operations and security operations management to this new class of entities.
Whether it is ultimately branded as an “agentic control plane” or an “agent ops management system” is, for Rajesh, less important than the underlying capability.
“When AI agents come into the technology infrastructure, how confident the CIOs can be, how confident the CSOs can be to allow the deployment and (management of) the operations is what we are thinking about. And this is what we will deliver for our customers.”
AI productivity, tempered with realism
Despite the uncertainties, Rajesh is personally bullish on AI’s productivity benefits.
Based on his own experience and that of his teams over the last three years, he said AI is already making him and the company “a lot more productive”. But he is equally clear that the industry is still in a “figuring out” phase.
“You have to accept what you know and also accept what you do not know,” he said. “Try to find the middle ground.”
That middle ground, for ManageEngine, seems to be about measured, governed adoption: embrace AI where it clearly adds value, but wrap it in controls that make sense to seasoned CIOs and CISOs.
A different stance on AI pricing
One stand out remark from Rajesh was his position on AI pricing. He reiterated more than once that ManageEngine does not intend to price its AI services separately.
While he did not elaborate on specific commercial models, the principle is clear: AI is to be embedded as part of the value customers already expect, not carved out as a premium add-on.
For customers, this could mean the ability to adopt AI-enhanced capabilities without negotiating separate AI line items.
In an environment where many vendors are still experimenting with surcharges and usage-based AI fees, this stance signals a deliberate attempt to frame AI as a natural evolution of ManageEngine’s platform, rather than a bolt-on product tier.
Extending the operations – security narrative to AI
Rajesh has already redefined ManageEngine’s own narrative from being “just” an IT operations management player to one that is equally focused on cybersecurity operations. He sees operations and security as complementary functions that must be tightly integrated, even if there is a “dotted line” between them.
The vision for an agentic control plane extends this convergence narrative into the AI era.
In Rajesh’s view, the enterprises that succeed with AI will not be those that adopt it fastest at any cost, but those that favour thoughtful, governed integration.