As companies across the world rush toward AI automation, a new concern is emerging – how to manage the fast-growing number of autonomous AI agents inside large organisations. A recent McKinsey study shows that nearly 65% of enterprises experimenting with generative or agentic AI don’t have a governance framework in place, leading to risks around data, security, and accountability.
Covasant, which works in the space of agentic AI governance and domain-specific automation, says this is where the next big shift in enterprise AI will happen. In this conversation, Srikanth Chakkilam, CEO & Executive Director, explains the company’s view on the changing landscape, the gaps they entered to solve, and why agent governance could soon become a mandatory enterprise layer.
Q1: What sparked the idea behind Covasant?
When AI agents first started entering enterprise workflows in 2023–24, we saw teams experimenting freely but without a clear structure. Companies were deploying agents for procurement, compliance and customer support – but there was no way to track, govern, or measure them.
It reminded us of the early cloud era, when teams used cloud tools freely before governance frameworks were created. That gap—between experimentation and safe, scalable production—led to the creation of Covasant.
Q2: In your early assessment, what was the biggest challenge for enterprises?
Fragmentation. Different teams were spinning up agents in silos. A recent Gartner note warned about “agent sprawl”—a situation where enterprises end up with dozens of autonomous agents with no central oversight.
This creates operational risk, data leak exposure, and inconsistent outcomes. We felt enterprises needed a domain-aware system to manage these agents safely and consistently.
Q3: Which industries are adopting Covasant’s tech the fastest?
Risk, compliance, fraud, and cybersecurity. These are high-stakes domains where errors can lead to fines or reputational loss. Globally, financial firms spent over $214 billion on compliance in 2023 (Thomson Reuters), so automation is becoming a priority.
Some of our agents process transactions for large clients, and the combined systems have handled over $1.4 trillion in value and helped detect anomalies worth around $500 million.
Q4: What are the biggest barriers enterprises face?
Four challenges appear repeatedly:
· Fragmented data systems
· Uncontrolled agent growth
· Pilots that don’t scale due to operational complexity
· Shortage of specialised talent
These are also reflected in O’Reilly’s 2024 AI survey, where 53% of companies said they lack the right talent to scale AI, and 38% said governance remains unclear. Our approach-domain-specific agents, unified AI fabric, and central governance—is aimed at closing these gaps.
