How it works

Every order is a set of promises. Keep them.

When a shopper places an order, your store has made three promises. The product promise: it looks, fits and performs the way the listing said. The delivery promise: it arrives when you said it would. The service promise: if something goes wrong, it is made right, fairly and fast. Revenue leaks when a promise breaks. Trust leaks at the same moment. So one idea serves both sides of the counter: help stores make true promises, and keep them.

The five leaks

Every rupee lost after checkout falls into one of five leaks.

A store's truth is scattered across tools that each see one slice. The returns app sees the reason the shopper picked. The courier sees scans. The reviews app sees sentiment. So a sizing problem gets blamed on the courier, and the same kurta keeps coming back. ALEAN's job is to put each lost rupee in the right leak.

An order's path from checkout to kept Orders travel a line from Checkout to Kept past five gates: Product, Promise, Payment, Discovery, Resolution. One order drips out at the Promise gate. The gate closes, and the next orders reach Kept. Checkout Kept Product Promise Payment Discovery Resolution
Promise is leaking: orders left the warehouse three days after purchase. Fixed: a dispatch-time alert. The next orders reach Kept. Illustration

Product

Fit, looks, quality. It didn’t match the listing.

Signals it reads
Return notes, reviews, pre-purchase questions
In the app today
The assistant answers size and fit questions from your product details and published Judge.me reviews.
Planned
Fit guidance on the product page, written from evidence (“Most buyers size up in this one”). Listing fixes ranked by rupees lost.

Promise

Dispatch or courier delay. It arrived after the date you gave.

Signals it reads
Dispatch time, courier scans against the promised date
In the app today
Order tracking for shoppers, with live courier updates when you connect Velocity.
Planned
Dispatch-time alerts, courier choice by pincode, honest delivery dates on the product page.

Payment

A cash-on-delivery order refused at the door, returned to origin.

Signals it reads
RTO tags, payment method, pincode history
In the app today
Not in the app yet.
Planned
Prepaid nudges and order confirmation where RTO risk is high.

Discovery

The shopper bought the wrong thing because a question went unanswered.

Signals it reads
Searches, zero-result queries, questions
In the app today
A shopping assistant that answers questions about products, sizes, delivery and your policies, in English or Hinglish.
Planned
Search tuning from what shoppers looked for and couldn’t find.

Resolution

A refund went out where an exchange would have kept the sale.

Signals it reads
Return outcomes, settlement type
In the app today
Exchange-first returns: the swap and store credit shown before the refund, under rules you set.
Planned
Outcome-aware suggestions per product: which items swap well, which don’t.

One loop

Sense, explain, act, prove, learn.

Today the widget and the returns inbox do the sensing and the acting. Explain, prove and learn are the next build, and they are what makes ALEAN an operating system rather than another returns app. Here is how the whole loop handles one product.

A worked example

One kurta through the loop

Illustration. The numbers are made up to show the method, not results.

A festive kurta has 40 returns in 60 days. Almost every return says “Delivery was late.” A returns app reports “Top reason: late delivery.” The merchant blames the courier and moves on.

  1. Sense. ALEAN pulls the 40 return lines from Shopify, the courier scans for each order, the dispatch timestamps, the product's reviews, and the searches that led shoppers to it.
  2. Explain. It checks the stated reason against what happened. 12 orders were late because they left the warehouse three days after purchase: a Promise leak, the store's own dispatch. 5 were late only on two pincodes served by one courier: Promise, courier routing. 23 arrived on time; their notes and reviews say “size chhota hai”: a Product leak. The kurta runs small. “Late” was just the first option on the form.
  3. Act. The product page tells shoppers “Most buyers size up in this one” and shows an honest delivery date for their pincode. The returns flow offers the next size up before offering a refund. Festive products get a dispatch-time alert.
  4. Prove. Sixty days later, against a holdout that didn't see the fit guidance, ALEAN reports how far the return rate fell and how much refund money stayed as exchanges. Every number labelled measured or estimated.
  5. Learn. Kurta, runs small, fit guidance: worked joins the cross-store outcomes, so the next brand that sells a kurta gets that fix suggested on day one.

The shopper never had to chat, fill a long form, or be surprised. The merchant never had to guess.

The receipt principle

A number you would defend to your accountant.

Every capability ALEAN ships must produce a measurable receipt: the change in Kept Revenue Rate per leak, returns and RTO prevented against a holdout, refunds turned into exchanges or store credit, the share of lost revenue with a known cause. A feature with no receipt does not ship. The dashboard is not a reporting feature; it is the product proving itself.

Receipts are part of the planned Brain. Today ALEAN records outcomes in Shopify; it does not yet report savings.

What we believe

  • Explain before acting. No fix without evidence it can trace to.
  • Shopify is the source of truth. ALEAN works from Shopify data alone. Connectors add depth; they are never a requirement.
  • Integrate, don't replace. If your tool does something well, we read from it and act through it.
  • AI is a tool, not the product. Rules and structured data first. Language models read language. No model ever sits in the money path.
  • Widgets render. They never decide.

The north star: a shopper is never surprised by what arrives, and a merchant never loses a rupee without knowing why.