MILLYA
MILLYA · in development
2026
The retail operating system

Recover stock.
Recover sales.
Buy better.

MILLYA is being built to connect what is happening across a retail business, explain what matters, prepare the next decision and remember whether it worked.

The interface gets simpler. The operating context underneath gets deeper. Human authority stays in control.

See the decision loop
Built from real retail operations
Evidence before recommendation
Human-approved consequential actions
Starts alongside existing systems
01The problem

Retail data is a mess.

Every retailer has data. The problem is seeing it properly. Sales live in one system. Stock in another. Buying in Excel. Finance with the accountant. Client notes in WhatsApp. Brand reports in manual files. To answer one basic question, teams export multiple ugly reports, combine them, and still end up guessing.

01

Sales reports that do not explain what changed

02

Stock reports that show problems too late

03

Buying files disconnected from sell-through

04

Client knowledge trapped in people's memory

05

Margin hidden until after the decision

06

Finance separated from daily retail reality

MILLYA fixes the first problem: it makes the business visible. Then it helps the team act.

02Decision loop

From operating data to a decision that can prove itself.

The goal is a closed management loop: understand the state, prepare the decision, keep authority explicit, then verify the real business result.

01

Know

Bring the relevant business state together without pretending missing or stale inputs are current truth.

02

Understand

Explain what changed, why it matters and where the risk or opportunity actually sits.

03

Compare

Look across credible options and trade-offs instead of optimizing one isolated KPI.

04

Prepare

Turn the recommendation into a reviewable next step — without silently turning advice into authority.

05

Approve

The authorized human keeps control of consequential decisions and execution.

06

Learn

Verify what actually happened, measure the outcome and carry the lesson into the next decision.

03The retail brain

Retail software records.
MILLYA is being built to reason.

The ambition is not another reporting layer. MILLYA is being built to connect the operating state of the business, explain the important change, prepare the next decision and remember whether that decision worked.

Consequential actions remain under human authority. Evidence comes before recommendation; outcome comes before learning.

01

Know

Bring the relevant operating context together across stock, sales, buying, clients, margin and finance.

02

Understand

Explain what changed, why it matters and what is still uncertain before a recommendation is made.

03

Decide

Compare the credible options against the business objective — not a single departmental KPI.

04

Learn

Remember the decision, verify what actually happened and improve the next recommendation from the outcome.

04The wedge

Start where retail loses money fastest.

Recover Stock → Recover Sales → Buy Better is the first focused value loop.

The views below are representative decision surfaces built with fabricated sample data. They explain product logic and interaction direction — they are not screenshots of the current production UI.

I · Recover stock

See stock risk while there is still time to act.

Ageing, location, sell-through and margin context come together so slow stock becomes a decision problem, not a season-end surprise.

Illustrative MILLYA view · stock/explorer
Sample data
On-hand
€1.96M
−4.2%
Sell-through 30d
62%
+3.1%
Avg margin
34.2%
+1.6%
Aged > 90d
€112K
−18%
Women / RTW / SS26
142 units78% ST31% margin42d
REORDER
Men / Outerwear / SS26
86 units31% ST18% margin91d
HOLD
Accessories / Leather
218 units91% ST44% margin18d
REORDER
Footwear / Sandals
64 units22% ST12% margin113d
MARKDOWN
4 of 6 sample lines shown
II · Recover sales

Turn inventory into a reason to contact the right client.

Client history and relevant product context become reviewable follow-up opportunities for the team — not another generic outreach list.

Illustrative MILLYA view · client/intel
Sample data

Today's recommended outreach

Ranked by margin opportunity
C-1042
Anchor
Last visit
12d ago
Spend 12m
€48.2K
LTV est.
€612K
RTWLeather
C-2117
Frequent
Last visit
4d ago
Spend 12m
€22.8K
LTV est.
€284K
BeautyGifting
C-0388
Anchor
Last visit
31d ago
Spend 12m
€94.1K
LTV est.
€1.1M
Fine Jewellery
C-3219
Rising
Last visit
2d ago
Spend 12m
€6.4K
LTV est.
€72K
Casual
Anchor: 142Frequent: 318Rising: 89
III · Buy better

Allocate the next euro with the business outcome in view.

Buying intelligence is being designed around gross profit, stock productivity, risk and optionality — including the possibility that the better decision is not to spend yet.

Illustrative MILLYA view · margin/engine
Sample data

Margin contribution by brand

30d
Brand 01
18%
Brand 02
14%
Brand 03
11%
Brand 04
9%
Brand 05
7%
Brand 06
6%
Brand 07
4%
Brand 08
3%

Prepared moves · human-approved

  • +€8.4KRecommended
    Reallocate floor space — Brand 04 → Brand 02
    Brand 02 shares space but contributes 1.5× margin.
  • +€3.2KRecommended
    Hold reorder — Brand 06 / Outerwear
    Sell-through 22% vs forecast 38%. Wait 14d.
  • −€1.1KRecommended
    Open markdown — Brand 07 / SS25 leftover
    Carrying cost crosses break-even on Jun 12.
05Ask MILLYA

Ask the business.
Then inspect the proof.

MILLYA is being designed so an owner can start with a natural question instead of hunting through dashboards. The answer stays connected to the evidence underneath it.

Conversation can explore broadly. Facts, approvals and execution remain governed separately.

Illustrative conversationSample data

Owner

Where is money getting stuck this week?

MILLYA

Two areas deserve attention: ageing stock in one location and a forward buying commitment whose recent sell-through is below plan. I would review the stock move first — it is lower risk and reversible.

Stock ageingSell-throughOpen commitments
Show evidence →
06Adoption

Start alongside what works. Earn the right to replace it.

Retailers do not change core systems because a pitch deck says they should. MILLYA's adoption path is being built around a lower-risk first step: connect to existing sources, prove value on real data and expand responsibility only when the evidence earns trust.

01

Connect read-only

Start from the systems already running the retailer. The first connection is designed to read and reconcile without writing back into the source.

02

Prove on real data

Show stock, sales and buying opportunities against the retailer's own history — with scope, freshness and evidence visible where it matters.

03

Expand only after trust

Operational responsibility grows deliberately. Existing systems do not need to be ripped out before MILLYA can start proving value.

The public promise is simple: first prove that MILLYA understands the business. Migration is a later decision, not the entry fee.

MILLYA Connect · read-only adoption layer
07Why it compounds

Features can be copied.
Operating memory is harder.

MILLYA's long-term defensibility is not a claim that one feature is impossible to reproduce. It is the compounding record of evidence, decisions, execution and outcomes around a retailer's real operating model.

Evidence compounds

The system keeps the business context behind important conclusions instead of reducing them to a chat answer or a chart screenshot.

Decisions accumulate

The useful memory is not conversation volume. It is what was decided, on what evidence, under which constraints and with what expected result.

Outcomes close the loop

A recommendation becomes more valuable when the system can verify execution and compare the real outcome with what was expected.

Context deepens

Each safely connected workflow gives MILLYA more governed operating context across the business — making the next decision less isolated.

Venture path

One wedge. A much larger operating layer.

Luxury and premium retail are the starting proof environment — not the ceiling of the architecture.

01 · Retailer

Operating system + intelligence for the retailer itself.

02 · Modules

Selected intelligence capabilities can sit on top of existing enterprise systems where full replacement is unrealistic.

03 · Network

Over time, permissioned retailer and brand workflows can create a broader intelligence layer around the same operating foundation.

08Origin proof

Built inside fashion retail. Designed for modern retail.

DiorBrunello CucinelliLoro PianaBottega VenetaThe RowGolden GooseZimmermannPhoebe PhiloKHAITE
09Founder-market fit

The advantage started before the software.

MILLYA is being built inside the operating reality it is meant to serve. Years of buying collections, carrying stock, running stores, managing clients and reconciling the numbers created the product brief before there was a software company around it.

Founder · Retail operator·

Ilya Loshak

Founder of Millya Group and operator of a multi-store retail business across buying, stock, clients, margin, finance and the daily floor. MILLYA grew from decisions the existing retail stack could record around — but could not help make.

Operating history → product thesis

  • 2011Millya Group is established in Cyprus.
  • 2012First Boutique opens its first store in Limassol.
  • 2020Years of operating across disconnected retail systems become the first product sketches.
  • 2026MILLYA is in active development as an AI-native retail operating system, starting from the highest-cost decision loops.
10Private conversation

The public site is the surface.The deeper story is private.

Retail operators can request a product walkthrough. Investors and strategic partners can request the private company narrative, product maturity context and deeper roadmap discussion. We intentionally do not publish the internal architecture or full product law here.