Insights matter only when they reflect real-world business consequences.
“You can’t fix what you don’t understand—and you can’t understand what you haven’t mapped. Observability must span tech and business, or it’s just noise.”~ Ananth Subramaniam, Sr. Executive Vice President & Head IT, Kotak AMC
At Kotak AMC, we operate in a world where milliseconds matter—and where one failed transaction could mean lost revenue, a missed NAV, or a customer lost to a competitor. In this environment, observability isn’t just a DevOps concern—it’s a business continuity imperative.
One of the biggest issues I’ve seen across organizations—and we’ve experienced it ourselves—is that technical observability without business context is incomplete. Yes, you can monitor infrastructure, trace APIs, and detect anomalies. But when something breaks, what does it mean for the business? Is it impacting a critical investor transaction? Is a distributor unable to complete a purchase? That’s the layer we need to surface.
We’ve moved toward a model where user journey observability is stitched end-to-end—from the front-end interaction, through the middleware and API layers, all the way to the RTA and back-office systems. This is especially critical for us because many of our touchpoints span third-party services, SaaS applications, and partner systems that we don’t fully control. If a single link in that chain is delayed or fails, the customer experience takes a hit—and our teams need to know exactly where and why.
We’ve also faced the challenge of RCA complexity. Often, issues manifest downstream, but the root cause lies upstream—in data mismatches, version misalignments, or middleware slowness. This is where AI-powered correlation across logs, traces, and events has helped, especially in pinning down problems that would otherwise require hours of manual effort and inter-team coordination.
And then there’s the human layer. I believe strongly that technical support teams must understand the business flows they’re safeguarding. A DevOps engineer might resolve an alert quickly—but without understanding how that alert maps to a failed fund switch or an unprocessed SIP, we miss the point. We’ve started sensitizing our support teams to the business processes, so they don’t just triage issues—they interpret them.
On the governance side, cost is a very real concern. With observability platforms typically priced per host, session, or agent, coverage can’t be blanket—it must be intentional. We’re increasingly focusing observability where the risk is highest: transaction-heavy flows, partner API chains, and investor-facing channels. At the same time, we keep iterating on dashboards that serve both tech and business audiences—so that we’re all looking at the same reality.
Ultimately, observability for us is about speed, precision, and alignment. Speed to detect and fix issues. Precision in knowing where and why. And alignment between IT and business so that RCA doesn’t just explain the past—it prevents future disruptions.