The problem
Serving millions of subscribers with a legacy support stack meant long wait times, costly phone and email handling, and no automation for routine requests. Agents needed a workspace that could absorb enormous volume — over a million tickets a month — without burying them, and that let AI take the repetitive work so people could focus on the conversations that mattered.
What I designed
- A high-volume agent workspace. A unified inbox and queue built for speed — surfacing the right conversation, context, and history so agents move fast across a relentless flow of tickets.
- AI + human, clearly divided. Made it obvious which requests AI was resolving and which needed a person, with clean handoffs so nothing fell through the cracks.
- Built for WhatsApp and 14 languages. Designed for the channel and language realities of an Indian consumer base at scale, keeping the experience consistent across them.
- Suggested responses & next-best actions. Surfaced agent-assist guidance during live conversations to cut resolution time on a high-volume operation.
Approach
The constraint that shaped everything was volume: every extra click costs real time when you multiply it by a million tickets. I designed for density without clutter, partnered with product and engineering on how AI and agents should share the work, and grounded decisions in how a real support floor operates day to day.
The operation, by the numbers
- 1M+Tickets / month on WhatsApp
- 14Languages, 24/7
- 5M+Routine requests handled by AI
Figures describe the TataPlay support operation (per Nextiva's published customer story), which integrated WhatsApp and AI to reduce resolution costs ~40%. My work focused on the agent portal experience within it. Full case study available on request.