Reimagining the AI Driven NPS System For Better Business Insights

Context

CleverTap’s NPS surveys capture rich qualitative feedback alongside scores. As adoption scaled, teams struggled to extract clear themes, sentiment, and actions from thousands of free-text comments.

This project focused on designing an AI-powered insights layer to make that feedback structured, explainable, and actionable at scale.

Problem

Teams faced three core challenges:

  • Qualitative overload — manual review doesn’t scale

  • Lack of actionable insight — dashboards show 'What' not 'Why'

  • Poor cross-team usability — insights are fragmented across functions

Design goal

Design a system that,

  • Converts unstructured comments into themes, sentiment, and summaries

  • Helps teams identify root causes, not just symptoms

  • Supports filtering, segmentation, and trends over time

  • Uses AI in a transparent, explainable way

  • Remains performant and cost-efficient at scale

Key design decisions

Design for cross-functional clarity

Outcome & Impact

While this case study focuses on the design phase, the intended impact is measurable:

  • Faster qualitative insight discovery

  • Reduced manual analysis effort

  • Clearer prioritization discussions

  • Better linkage between customer feedback and product decisions

  • Scales to ~800+ GB of long-term data

  • Balances performance, cost, and value delivered

Learnings (Redo the content)

This project reinforced that good AI UX is constraint-aware UX.

The most important design decisions were not visual:

  • When not to recompute

  • How to communicate freshness and confidence

  • Where to stop automation and return control to the user

Designing AI responsibly meant designing clarity, trust, and restraint.