Product case study · Varsee

From an idea with no documentation to an investor‑ready product in 6 weeks.

Varsee connects online shoppers with a store's own salespeople through live one-to-one video, with AI that reads what the customer wants, suggests products and coaches the salesperson.

ClientVarsee (Boomdevs venture)
RoleCo-founder & Head of Design
TeamFounding team of four
TimelineFrom 10 Sep 2025 · prototype in 6 weeks
StatusPrototype · pitching to investors
TL;DR
Challenge
A store owner showed products to customers over FaceTime and converted over 80% of them, compared with 2–3% on the website. Varsee had to turn that into a product, starting from an idea with no documentation or plan.
My role
Co-founder and Head of Design. I owned the product: the problem framing, scope, the AI layer, the interaction model, pricing and every product decision. I also designed all five surfaces.
Result
A full product prototype in six weeks, covering every phase of the roadmap, which the founders now use to pitch investors. Build starts after funding.
6wk
From idea to full prototype
5
Product surfaces
3
User types: shoppers, salespeople, store owners
01 · Key decisions

One-to-one video, AI support and no app to install

Starting with one-to-one live video

Problem
Online catalogs convert 2–3%; shoppers can't see or ask.
Decision
One-to-one live video between the shopper and the store's own staff, aimed at small boutiques.
Trade-off
No reach to thousands of viewers per stream; every session needs a salesperson's time.
Said no to
Influencer-style one-to-many livestreams.
Worth it because
It copies the behaviour that actually converted: a real person showing a real product.

AI beside the salesperson

Problem
AI could run the whole sale, but trust is what online shopping lacks.
Decision
AI supports the salesperson in real time (it reads intent, suggests products and coaches) and gives shoppers AI try-on, during a session or on their own.
Trade-off
Harder than a chatbot, and a person still has to be online.
Said no to
An AI shopping bot replacing the human.
Worth it because
The human builds trust; the AI makes every salesperson better.

Easy for stores to adopt

Problem
Small retailers won't rebuild their store or ask shoppers to install an app.
Decision
Browser-based, no app: a “Shop Live” button on any product page, and a shared cart that hands off to the retailer's existing checkout, set up in minutes.
Trade-off
Varsee depends on other platforms' checkouts.
Said no to
Its own app, its own checkout or a marketplace.
Worth it because
A store can start the same day without changing how it sells.
02 · Execution

From an idea to five surfaces in six weeks

Starting on 10 September 2025 with only an idea, I wrote the product definition, scoped it, and designed five surfaces in six weeks: the marketing site, the customer experience (Shop Live, video session, AI try-on), the salesperson's mobile app, the store owner's web app (dashboard, settings, reports) and the system admin.

Live video session: a salesperson shows a jacket, with product details and Try-on and Buy Now buttons
Customer · live one-to-one session with a store salesperson
AI virtual try-on result next to the shopper's original photo
AI try-on inside the session

For salespeople streaming from mobile, desktop or smart glasses, I defined the gaze, gesture and overlay patterns in workshops with Varsee's CTO and lead developer.

Salesperson mobile app home with availability toggle and session stats
Salesperson app · go online, today's sessions and ratings

I designed the whole product, every phase, so investors could see the full vision, and drew a clear build line: the first release is live sessions, intent and product suggestions; AI coaching and try-on follow in a later phase.

First release Live sessions
  • Live one-to-one sessions
  • Intent
  • Product suggestions
Later phase AI coaching and try-on
  • AI coaching
  • AI try-on
Build line · designed in full, built in phases

Pricing is per-user SaaS in three tiers, so cost scales with the sales team, not with traffic.

Store owner dashboard with revenue, conversion, sessions and satisfaction
Store owner dashboard · sample data
03 · Outcome

A full prototype in six weeks

  • A full product prototype in 6 weeks, from an idea with no documentation or plan.
  • 5 surfaces designed, covering every phase of the roadmap.
  • The founders now use it to pitch investors. Build starts after funding.
04 · What I'd do differently

Test the AI before every screen

I designed every phase so investors could see the whole product. Next time I'd test the AI layer first (intent detection and product suggestions on real recorded sales calls), because that's where Varsee succeeds or fails.