Returns and unsold stock, sold again.

Brands send Birl their returned and unsold clothes. We check each one, fix what is worth fixing, and sell it again. Our software runs every step.

Since June 2026: 4,000+ items received, £78k resold, 3,600+ live for sale today, on a platform that costs $5 (about £4) a month to run.

AYM Ava Reversible Maxi Dress in green, worn by a model
The dress's real swing tag, as the warehouse phone sees it
AYM · Ava Reversible Maxi Dress

A phone reads the label. The system knows the rest.

The dress comes back as a return. A warehouse phone scans its tag, and Birl finds the exact product and size in AYM's catalogue, plus how many are already on sale.

  1. Pick the job: triage and grade
  2. Scan the tag
  3. Ava Reversible Maxi Dress found, with live stock

9 in 10 items are identified without anyone searching by hand. Staff confirm any match the system is unsure about.

Starting…
The real warehouse app · demo data

Is it worth fixing? Birl decides in a second.

A seam has come loose. Staff mark it as needing sewing, and Birl weighs what the dress will sell for against the fees and the cost of the repair.

  1. Condition: needs sewing
  2. The fix-or-not engine: worth restoring

527 fix-or-not decisions since late September. Items not worth fixing are set aside, so no labour is wasted, and every item is told which box it goes in.

Starting…
The real warehouse app · demo data

One listing. Every place it can sell.

Photos, price and stock go to AYM's own resale site, eBay, Vestiaire and Depop at once. When it sells anywhere, it comes down everywhere else.

  1. The stock queue
  2. The Ava: photos and details
  3. Where it's live

3,647 items live across four channels. AI can create model photos from one shot; staff approve them, and they never hide a fault.

Starting…
The real Birl dashboard · demo data

Sold, packed and answered.

Orders from every channel land in one queue with the label ready. Questions land in one inbox, where AI answers the simple ones straight away.

  1. Every order, every channel, one queue
  2. A buyer asks about the Ava
  3. AI answers from live stock, in seconds

Offers, returns, refunds and complaints always go to a person. Every new answer we write can become a house answer.

Starting…
The real Birl dashboard · invented customers

Every sale makes the next decision better.

One record follows the dress, and every item like it, from arrival to sale, return and brand report. That record shows what each item really sold for after repair and fees.

Items sell in 8 days on own sites and 12 days on eBay, and 92% arrive ready to sell.

Our best month so far was 1,707 items (July). The plan for December 2027 needs about 2.5 times that. Recent months were limited by how much stock came in, not by the system.

ReceivedSold
Measured in Birl's systems. June is a part month; October covers 1–20.

Pull back: everything that dress passed through.

Each piece is built by Birl and runs from one dashboard, on a platform that costs $5 (about £4) a month. Our money goes into engineering, not servers.

  • Birl photo matching

    Knows a garment from a photo or its label. 9 in 10 need no manual search.

  • Fix-or-not engine

    Decides if a repair pays, from real sale prices and real repair costs.

  • Birl Hub

    The warehouse computer that prints tags and keeps marketplaces in sync.

  • AI product photos

    Model photos from a single shot, approved by staff before they go out.

The AYM Ava Reversible Maxi Dress
AYM · Ava Reversible Maxi Dress

  • Sell everywhere at once

    One listing to brand sites, eBay, Vestiaire and Depop. Sold once, removed everywhere.

  • AI customer service

    One inbox for every channel. AI answers the simple questions; people handle the rest.

  • Ops assistant

    Staff ask in plain English. The AI prepares it; a person clicks to approve.

  • Brand reporting

    Each brand sees its sales, condition, time to sell and an environmental report.

What the investment brings to market.

Three ways to grow, each on the platform that already runs every day.

Trade-in 2.0

Credit that matches what an item will really sell for.

Trade-in has run since November 2024: 868 trade-ins for 17 brands and clubs. Version 2.0 is built. Next, the offer depends on what similar items really sold for, their condition and how fast the style sells, paid straight into the shopper's store account.

Done when: a large brand runs it, with offers checked against real sale prices for 100+ items.

Built · dynamic credit in 2027

Peer-to-peer

Shoppers resell on the brand's own site.

Most of it is already built, and payments run on Stripe. The investment finishes the brand integration and takes it to a pilot in mid-2027. It works for brands on any platform.

Mostly built · pilot mid-2027

Birl platform licence

Other companies run their own resale on Birl.

A Birl engineer sets it up and trains their team; then they run it on their own. Low running costs make it high-margin.

Opportunity

The next 12 months, in priority order.

Strengthen the core Grow the platform New, if conditions are met Pilot milestone
Mid-2027Pilots running; first partner region
Dec 202713 brands, about 4,300 items a month
Jul 2028Second region, reusing what works
Dec 202828 brands, about 12,700 items a month

What could go wrong, and what we do about it.

RiskSafeguard todayNext step
Volume grows faster than the systemCapacity grows automatically; the database is at 3% of its limit.Stress test at twice the 2027 target
One brand sees another's dataEvery record and file is tied to one brand; brand users see only their own.Automatic checks that prove it
Data loss or outageRestore to any minute in the last 30 days; nightly stock and sales checks with alerts.Regular separate backups; a recovery plan
A bad updateEvery change is tried on a full copy of live data first.Every automatic test on every release
AI makes a mistakePeople approve anything that moves money or appears in front of buyers.Measure accuracy by brand
AI costs rise with volumePay per use; covered by startup credits today.Track AI cost per item