0-to-1 AI Product · Built with Lovable

PULSE — AI Customer Risk Radar

An AI-powered Customer Risk Radar for Customer Success teams — it surfaces which accounts are healthy, need attention, or are at high risk, and tells you why, so you can act before churn. Includes an overview dashboard, Customer 360 profiles, explainable risk scoring, ticket intelligence, an action center with suggested outreach, and a Chrome extension that overlays risk scores directly onto your ticketing app or CRM.

Role
Solo builder — problem to working prototype
Method
Explainable scoring design, prompt engineering
Scope
Customer Success, account risk radar
The problem

A risk score nobody can explain is a risk score nobody trusts

Health scores exist in most CS tools, but when a CSM can't see why an account dropped from green to red, they either ignore the score or spend as much time investigating it as they would have without one. I wanted risk scoring that shows its work.

Solution

PULSE: explainable risk, surfaced before it becomes churn

RADAR

Healthy, attention, or at-risk — at a glance

An overview dashboard and Customer 360 profiles sort every account into a clear status, backed by the signals behind it, not a single opaque number.

EXPLAINABLE SCORING

Every score shows its reasoning

Risk scores are broken down into the usage drops, ticket patterns and champion activity that produced them — so a CSM can see why, not just what.

ACTION + EXTENSION

Suggested outreach, and a CRM overlay

An action center suggests next steps directly from the risk breakdown, and a Chrome extension overlays the score onto your ticketing app or CRM — no extra tab required.

Prototype

A working risk radar, not a black-box score

PULSE is built and live — the dashboard, Customer 360 profiles, explainable scoring and action center all work end to end in the prototype below.

Live prototype — PULSE Open full screen ↗
Open interactive prototype ↗
Process

Built on a Quarterly Retention Program that actually cut churn 35%

Discovery
Problem discovery — started from watching CSMs distrust an unexplainable health score at Zoho, and design around what made scoring trustworthy there.
Flow
Scoring architecture — designed the breakdown from usage, tickets and champion activity into one explainable score.
Prompts
Prompt engineering — wrote the prompts that turn a risk breakdown into suggested outreach and a meeting-ready narrative.
Output
Build — shipped a working, clickable radar on Lovable, plus a Chrome extension concept for CRM overlay.