CASE STUDY
PRODUCTAI

Your customers speak. Hear what they say.

CLIENTPercept Pulse
INDUSTRYCustomer Experience
TECHNOLOGYAI Web Application
Percept Pulse ▪ Your customers speak. Hear what they say.

PERCEPT PULSE

OVERVIEW

Percept Pulse needed to turn a universal, low-effort customer reaction (an emoji tap, or a few spoken words) into structured, actionable sentiment data, across every branch and every language, without asking a customer to fill out a survey they'll never finish.

ARCHITECTURE

The AI pipeline is what earns the product its name: a voice recording is transcribed, language-detected, translated to English, and classified into a five-point sentiment scale (angry to excited) before the tenant's dashboard ever shows the result. Every response is tagged to a branch and a staff member at the point of submission, so the live dashboard can roll sentiment up by location or by person without a separate analytics step.

KEY DECISIONS

Voice as a first-class input

Speech is transcribed, auto-translated, and sentiment-classified in one pipeline, so a customer can speak in Urdu or French and the manager reading the dashboard sees an English summary and a sentiment score, not a language they need to translate themselves.

Emoji-first, five-point scale

The primary interaction is a five-emoji tap (angry to excited) that takes under three seconds, deliberately not a text survey, because response rate, not survey depth, was the metric that mattered for a device sitting on a counter.

Branch and staff attribution from day one

Every piece of feedback carries a location and staff-member tag from the moment it's submitted, so leaderboards and per-location trends are a query, not a manual reconciliation.

Installable, offline, and never silent

Both the manager dashboard and the staff portal install like a native app with no app-store step, keep working when a connection drops, and push an instant notification the moment angry or sad feedback comes in, even with the app closed.

TECH STACK
ReactNode.jsSpeech-to-TextMachine TranslationSentiment ClassificationProgressive Web AppPush Notifications
THE OUTCOME

Businesses get a live, language-agnostic read on customer sentiment per branch and per staff member, from a device that takes a customer under five seconds to use, with negative feedback pushed to a manager's device in real time instead of surfacing in a monthly report nobody reads in time to act on it.

Have a project in mind? Let's talk.

START A CONVERSATION
TRY IT LIVE

Edit the code, watch it run.

Change the function on the left: your letter's background updates instantly. No reload, no waiting.

highlight.js
OUTPUT: A..Z, click a letter and watch the alphabet drop
ABCDEFGHIJKLMNOPQRSTUVWXYZ
HELLO@20THLETTER.COM LET'S SET IT IN TYPE HELLO@20THLETTER.COM LET'S SET IT IN TYPE HELLO@20THLETTER.COM LET'S SET IT IN TYPE HELLO@20THLETTER.COM LET'S SET IT IN TYPE