Acko Insurance: Reducing 20% Opex using the observability platform

Acko General Insurance, launched in 2016, is India’s first fully digital, direct-to-consumer (D2C) insurance company. The digital only General Insurance company serves over 200 million families across the country. With no intermediaries in its business model, every customer interaction — from buying a policy to raising a claim — happens directly through Acko’s website, app, or support channels. This digital-first approach makes system reliability, performance, and customer experience absolutely critical. 

To maintain seamless operations at scale and ensure uninterrupted service during peak traffic periods, Acko has implemented New Relic’s observability platform — enabling real-time visibility, faster incident response, and data-driven optimization across its digital ecosystem.

Acko’s Atypical Observability Demands

Acko faces distinct business and technical challenges as a digital-first insurer. The company must manage a vast spectrum of consumer behaviors and product complexities—from seamless, one-rupee ride insurance embedded in a ride-sharing app, to multi-week buying cycles for complex health and life insurance policies.

This demands an incredibly flexible and unified information architecture that can cater to instantaneous transactions and prolonged research journeys simultaneously. Technologically, Acko operates as a hybrid entity: requiring the agility and responsiveness of an e-commerce platform, while also maintaining the long-term data integrity and management, typical of a financial services company. 

This duality means Acko’s observability requirements are distinct from traditional e-commerce or insurance players. Further complicating this is their multi-cloud environment and the development of their proprietary insurance operating system platform, sureOS, that Acko itself builds and consumes. These factors collectively create an inherently complex system where traditional, fragmented monitoring solutions simply cannot provide the holistic, real-time insights needed to ensure continuous availability and performance.

“Our systems are at the center of our product strategy. So, observability has been our engineering focus since day one of operation,” said Harish Rama Rao, Senior Vice President of Engineering at Acko. 

Observability Powers Cloud, Human Resource Rationalisation

The Observability solution provides a complete overview of the health and performance of the application, network, infrastructure and users. Kind of acts as a vantage point for the IT teams to keep tabs on the system vitals and in case of bottlenecks - take remedial action in real time.

The major difference between the earlier platform and the current observability is the aspect of being ‘Real-Time’. The ‘Infrastructure as code’ like feature tracks in real time the various services and components, helping detect anomalies and slowdowns within seconds.

“By ensuring we are always finely tuned to meet the consumer demand during peak sales seasons, we are able to save costs from the earlier times. For example, similar applications running from other organisations, you will say, you're running at X load, or there is a peak season coming in at 1.5X. So there's no rationale without Application Performance Monitoring (APM is a component of the observability platform) to go 1.5X ?. We have metrics or emitters, built in as code. In a way, think about it like infrastructure as code. Those metrics get emitted naturally, we consume them. Now, the New Relic (Observability platform) allows Service Level Objectives (SLOs) to be written as terraforms (Tool used by Cloud Engineers) that ensures that the systems automatically respond to change in our demand, which is change in our scaling parameters,” explained Rama Rao.

Acko has been able to achieve considerable amount of savings after using the platform, “We have been able to save between 15 to 20 percent, by optimising the machines running at optimal values of throughput, optimal values of uptime, and optimal values of availability from our previous benchmarks, where it was more pre-calculated methods of scaling,” says Rama Rao.

Another major advantage of the platform is during major upgrades, new feature deployments, CI/CD cycles, etc. In reducing the days for which old systems are kept live during red-green deployments. Usually the older systems are kept on while new features or upgrades are deployed as a back-up. In case the new system fails, the old takes over however there are costs involved.

Rama Rao explains the advantage of the platform in reducing the mean time to detect that something is wrong, “We use the telemetry and the APM system to do synthetic testing. So if we had to run two parallel systems for about a month, now with this right way of engineering, we are able to cut it down to four or five days. Hence, naturally, there are a lot of savings, in terms of our ability to change, ability to move, ability to sort of morph into the newer platform technology.”

Real World Use Cases

During Diwali, Acko Insurance experiences a sharp surge in traffic as customers renew or purchase motor insurance before the holidays; however the technology team can predict and monitor load patterns across services in real time.

As a D2C insurer, Acko has no intermediaries—its website, mobile app, and support centre are the only customer touchpoints. Observability is embedded across all these channels to ensure a seamless experience. In the contact centre, the platform tracks - Response time to customer calls; Average handling time per issue, etc.

Using the observability platform, Acko has recently launched an AI-driven product that analyses customer calls end-to-end. It converts call recordings from speech to text to assess whether a customer’s issue was resolved, benchmarks agent performance, and automatically pushes tailored training material to agents before their next call. Extending the use of the observability platform beyond system metrics to human interactions—for instance, analysing whether a call drop was accidental or intentional.

Acko goes beyond uptime monitoring and analyse system performance from both consumer and internal perspectives. By correlating service usage data with infrastructure consumption, the team optimized cloud utilization, delivering nearly 20% reduction in operational costs.

Acko has ensured 24/7 system uptime and efficient scaling of its insurance products. The platform has resulted in a 20% reduction in costs, a 20% reduction in Mean Time to Resolution (MTTR), and Mean Time to Detect (MTTD) reducing to under five minutes. It has also enabled Acko to accelerate software releases.

Future plans

Acko wants to leverage the Agentic AI tools from the observability platform. Currently, Acko’s IT team still has to manually set up and manage many parts of their monitoring system. “Today it is still a two to three step process. There is writing the metric or else embedding the metric, the telemetry; the alerting, alarming, or the metrics getting emitted; dashboarding and alerting along with it. Then there is observing itself,” informs Rama Rao. 

In the future, Acko wants to rely on smart ‘agents’ built into the observability tool that can automatically understand what’s happening, fix simple issues on their own, and only call humans for complex problems. That way, the team doesn’t waste time or sleep on routine maintenance work. Rama Rao says, “When agenting systems get developed by the tool provider, a lot of this is automated.”

Acko wants to make its software release process (called CI/CD — Continuous Integration and Continuous Deployment) much more automated and reliable using observability and performance monitoring tools.

Right now, when they release new updates, people manually watch the systems to check if everything is working fine. If something fails, they step in to fix or roll back the changes. Acko wants tools to do this automatically — for example, if a new update fails (a ‘red’ deployment), the system should instantly roll back to the older version (‘blue’) without waiting for humans to intervene.

They also want to define all their service goals and reliability rules (SLOs) directly in code — instead of keeping them in documents that engineers must manually read and follow. This way, everything becomes transparent, version-controlled, and auditable — every engineer can see the exact settings, make changes through code commits, and review them just like normal software.