Evergent cuts call resolution time by 50% with AI-powered, context-aware platform

Evergent supports media and entertainment platforms in managing subscriber lifecycle operations, including billing, subscriptions, and customer interactions. These platforms rely on customer support teams to handle a high volume of queries across multiple channels, making timely access to accurate subscriber data critical. 

In many OTT environments, customer support operations are distributed across multiple systems. Customer care, billing, analytics, payments, and content platforms are often managed by different vendors, each with its own interface and data structure. As a result, human customer support agents must navigate several disconnected applications to resolve a single query.

In one deployment, teams operated across six different applications, creating what Vivek Sadhineni, Product Manager at Evergent, described as a “broken pipeline,” where performance depended on whether integrations were functioning as expected.

This fragmentation limited access to real-time subscriber context and created a disjointed customer experience. Even logged-in users were asked to provide basic details. As Sadhineni explains, “As a logged-in user, I expect it to recognize me and get me the information that I want.” At the same time, rule-based support systems lacked the ability to interpret user intent or personalize responses, while interactions from non-logged-in users were not consistently captured, reducing visibility and follow-up opportunities.

Evergent saw the unification of fragmented systems as an opportunity to centralize subscriber data and deliver more contextual, efficient customer support while reducing handling time and operational complexity.

To address fragmented customer support workflows, Evergent developed ACE, an AI-powered platform that unifies customer interactions with subscriber lifecycle data to enable context-aware support experiences.

“We set out to build a context-aware digital experience where the system understands the customer, their subscription, and their interactions, and responds in a more natural, conversational way,” says Ria Kapila, Chief AI and Product Officer at Evergent.

In the initial phase, Evergent replaced an existing CRM-based communication system and brought customer interactions into its platform using Amazon Connect. This reduced dependence on multiple systems and minimized the need for agents to switch between applications, allowing them to access subscriber, billing, and order information within a single workflow.

"We built a platform that understands who the customer is, what they're subscribed to, and what they've already done, so the agent doesn't have to ask," says Sadhineni. The platform integrates directly with Evergent’s subscriber lifecycle management system, which acts as the source of truth for subscriber, payment, and order data. This enables context-aware interactions, where logged-in users are recognized and relevant information is retrieved automatically.

To enable more intelligent, conversational experiences, Evergent used Amazon Bedrock to generate context-aware responses based on real-time subscriber data, moving beyond static, rule-based workflows. Amazon Lex is used to interpret customer intent and guide interactions dynamically, enabling more natural and responsive conversations.

By consolidating fragmented systems and introducing AI-driven capabilities, Evergent improved the efficiency of its customer support workflows. “With AI capabilities on Amazon Bedrock, handling time reduced by approximately 40 percent, while time to resolution decreased by up to 50 percent because agents can access the right information more quickly,” says Kapila.

Sadhineni adds, “Bringing customer information and workflows together in a single interface helps agents work more efficiently and deliver a more consistent support experience.”

The platform also reduced the need for agents to switch between multiple systems, saving time at each step of the interaction. Previously, navigating across multiple screens could take up to 40 seconds per query, which has now been significantly reduced through a unified workflow. 
In addition, the platform enhanced visibility into customer interactions by capturing data across support channels, including interactions from non-logged-in users, providing a more complete view of customer behaviour and enabling better follow-up and engagement.


As a next steps, Evergent is extending the platform with an intelligent agent workspace designed to capture conversation context, summarize interactions, and recommend next-best actions to improve response speed and consistency.