Business Problem

• 20 million new customer voices go online every day. Brands use almost none of it.

• Existing tools told teams sentiment is down never why. One blended score, no cause.

• One report served every department, so it served none. Marketing, R&D, Sales and CX all needed different cuts of the same data.

• The intelligence existed in the MASI engine. It just wasn't legible to the people who had to act on it.

• Design question: How do you take aspect-level sentiment across 100+ categories and 100M+ voices, and make it usable by a marketing manager at 9am no analyst in the loop?

Design Impact

• Department-first IA. Replaced the single shared dashboard with role-scoped views. Marketing sees keywords, R&D sees product gaps, Sales gets battle cards, CX sees complaint spikes. Same engine, five different front doors.

• Aspect over average. Designed the core visual pattern around independently scored aspects (battery, comfort, dining, value) instead of one composite number, so a brand can read "excellent on location, poor on value" at a glance.

• Segment layering. Buyer-type filters (first-timer vs upgrade, business vs family traveller) built into the primary view, not buried in settings.

• SmaartAnalyst. Designed a plain-language query surface on top of structured MASI output — grounded answers, suggested prompt starters to solve the blank-input problem, and visible data provenance to build trust in an AI answer.

• Multilingual by default. Interface and query flow designed for English, Hindi, Tamil and Telugu without layout breakage.

• Outcome: Deployed across automobiles, hotels, electronics, FMCG and D2C. At Mars Petcare, the platform consolidated fragmented review sources into one portal and improved the CRM team's response capability — surfacing product strengths and gaps their existing tooling had flattened.