The whole thing, end to end

One request. Six steps. Both sides of the counter.

Chapter 2 // Next is two connected surfaces: the shopper's side, where a sentence becomes a styled, in-stock, correctly sized look, and the operator's side, where the same activity becomes demand, approvals and decisions. This page walks the same afternoon from both angles.

Pieces live

35

Titles, prices and imagery from Chapter 2's storefront

Stores

3

Khar West and Colaba in Mumbai, Basant Lok in Delhi

Demo shoppers

25

With orders, fit profiles and conversations

Invented promises

0

Stock, sizes, prices and policy come from the database

01

Shopper

She describes an evening, not a product

“Cousin's engagement in Bandra on Saturday. All black, nothing shiny, under twelve thousand.” The stylist reads the occasion, the colour, the budget and the city and builds complete looks from pieces that are actually in stock.

Try the stylist

What the operator sees

The prompt, the parsed intent and the looks shown are all logged. The operator can see what people are asking for in their own words.

02

Shopper

She asks the only question that matters: will it fit

Height, usual size, how she likes things to sit. The advisor answers with one size, a reason, and an honest confidence level. Where the size chart is missing it says so instead of guessing.

Check a size

What the operator sees

Every low-confidence answer points at a missing measurement, so the catalogue team knows exactly which piece to measure next.

03

Shopper

Her size is sold out, so she joins the list

No dead end. She leaves her size, her city and permission to be told once when it returns. Nothing is promised and no date is invented.

See a sold-out piece

What the operator sees

Demand stacks up by piece, size and city with a value attached, so restocking becomes an ordered list instead of a hunch.

04

Shopper

Or she holds it at Khar West and tries it on

If the size is on a store floor she holds it for 48 hours and walks in. Online interest becomes a real fitting-room appointment.

Browse store stock

What the operator sees

Store managers see the hold, mark it collected or release it, and the unit goes straight back into available stock.

05

Operator

The assistant drafts the reply, and knows when to stop

Returns, exchanges, delivery dates and anything high-risk get drafted but never sent alone. If two policies contradict each other, the topic is blocked until a human decides which wording is true.

Operator view

What the operator sees

Confidence, risk and the reason for stopping are written down for every draft — including when the model was unavailable and written rules answered instead.

06

Operator

The week ends with a decision, not a dashboard

What to restock, in which sizes, for which city; which policy to publish; which drop to plan. All counted from orders, waitlists, holds and returns.

Back to the stylist

What the operator sees

Assisted revenue, size-return rate and reply times sit next to the demand list, so the trade-off is visible in one place.

Honesty by design

What this build refuses to do

A demo that fakes the hard parts can't be trusted with the easy ones.

  • — Never claims stock, a size, a price or a delivery date that isn't in the database.
  • — Never sends a message on a channel a shopper didn't agree to.
  • — Never answers a policy question while two policies contradict each other.
  • — Never hides a fallback: if the model is unavailable, the answer says so.
  • — WhatsApp, Instagram and payments are shown as not connected, because they aren't.

Where to start

Walk it yourself

The operator side asks you to sign in and pick a seat. The first person to sign in can take the owner seat; other seats see less, on purpose.