INTERACTIVE PRODUCT VISIONIllustrative data used to demonstrate the intended Departmint workflow.

See how a fresh decision gets made.

A simulated store, a real workflow. Follow one category from signal to recommendation to economic outcome — and choose what you would do at the decision point.

DATARISKDECISIONACTIONECONOMIC OUTCOME
DEPARTMINTFRESHFRESH PRODUCE · CONTROL ROOM
DEMO STORE — DUBAI MARINASIMULATED ENVIRONMENT
FRESH INVENTORY VALUE
AED 28,460
REQUIRING ATTENTION
AED 4,280
POTENTIAL VALUE AT RISK
AED 1,840
ACTIVE RECOMMENDATIONS
7
NEED INSPECTION
4
MARKDOWN OPPORTUNITIES
3
01 — CATEGORY OVERVIEW

Six categories. Departmint ranks where attention is needed.

Select a category to see its current state. Strawberries carry the full walkthrough below.
SELECTED
Strawberries
INVENTORY
52 boxes
DAYS OF COVER
4.7 days
RECOMMENDED ACTION
Review markdown
PILOT SUITABILITY
9 / 10
02 — THE SCENARIO

Strawberries

BATCH ST-240812RECEIVED 10 AUGDUBAI MARINA · FRESH PRODUCE
SIMULATED TIME
12 AUG — 14:00
INVENTORY
52 boxes
CURRENT PRICE
AED 18.95
SALES VELOCITY
11 /day
SOLD TODAY
4 boxes
CURRENT MARKDOWN
0%
WASTE, LAST CYCLE
6 boxes
CONDITION TREND
Declining
RISK LEVEL
HIGH

Two kinds of signal, combined into one decision.

SIMULATED — CONDITION MANUALLY OBSERVED
COMMERCIAL SIGNALS
What the retail system already holds.
Inventory52 boxes
Sales velocity11 /day
PriceAED 18.95
Demand trendDeclining
CONDITION SIGNALS
What the shelf shows. Observed by hand in the pilot, not sensed.
Observed conditionDeclining
Visible anomalies3
Temperature exposureNormal
CONDITION DISTRIBUTION
The 52 boxes are not in one state.
41good
8sell soon
3require review
DEPARTMINT DECISION —
REVIEW INTERVENTION

Condition is an input, not the product. The output is still a commercial decision, and the test is whether adding the left-hand column changes it enough to be worth collecting.

03 — THE DECISION WINDOW

The store's usual markdown happens at 18:00. That may be too late.

Departmint estimates that waiting until the evening markdown reduces the intervention window — and the pricing flexibility still available inside it.

TOO EARLY
Unnecessary margin loss
DECISION WINDOW
Intervention still recovers value
TOO LATE
Deeper markdown or waste
NOW — 12 AUG — 14:00
08:00
Stable
12:00
Monitor
14:00
Current state
18:00
Usual markdown
TOMORROW 08:00
High risk
04 — EXPLAINABILITY

Why is this inventory being flagged?

Every recommendation shows its inputs. A store manager who cannot see the reasoning has no basis to disagree with it — and disagreement is information we need.

DEPARTMINT ASSESSMENT
Current inventory may exceed expected full-price demand before the next normal intervention point.
Inventory remaining52 boxes
Current velocity11 / day
Recent velocitySLOWING
Product observationsTRENDING DOWN
Historical markdown timingLATE IN DAY
Expected sell-throughBELOW EXPOSURE
05 — THE RECOMMENDATION · YOUR DECISION
RECOMMENDED ACTION
Review markdown now
AED 18.95→AED 14.95
RECOMMENDED REVIEW WINDOW
Within 2 hours
REASON
Earlier intervention may increase sell-through before additional value is lost.
DECISION SUPPORT — DEPARTMINT DOES NOT
EXECUTE THE MARKDOWN. THE STORE DOES.
WHAT WOULD YOU DO?
OUTCOME OF YOUR CHOICE
Choose an action to see the simulated economic consequence, compared against the store’s current process.
07 — THE ECONOMIC LOGIC

A good decision is not the one with the lowest waste.

Waste can be driven to almost zero by discounting everything, early and deeply. That is not a win. This is the equation Departmint has to satisfy instead.

RECOVERED
Margin recovered
COST
Markdown cost
COST
Waste cost
COST
Staff & operating cost
COST
Unnecessary interventions
=Incremental economic value
08 — WHEN DEPARTMINT IS WRONG

A false positive is a cost, and we count it.

A system that sends staff to inspect healthy product, or discounts stock that would have sold, destroys value while looking accurate. Departmint measures the cost of unnecessary interventions alongside the successful ones.

Try it: this avocado flag is a false positive. Inspect it.

Avocados
FLAGGED — POTENTIAL RISK
Departmint flags potential deterioration risk across 84 units and requests a visual check before any action.
Logged asPENDING INSPECTION
09 — ONE SYSTEM, DIFFERENT DECISIONS

The same data, answering a different question for each role.

STORE OPERATIONS
What should the team do now?

On the floor the interface is a task list, not a dashboard. One instruction, the reason behind it, and two ways to respond. Nothing else.

Store actions4
HIGH PRIORITY
Strawberries
Review markdown
AED 18.95 → 14.95
INSPECTION
Tomatoes — Shelf 04
Inspect 4 units
MONITOR
Avocados
No action · recheck in 4h
COMPLETED
Blueberries
Inspected · no intervention
10 — WHAT A REAL PILOT WOULD TEST

Everything above is a simulation. This is what we would actually measure.

THE RETAILER'S EXISTING
PROCESS IS THE CONTROL
OPERATIONAL VALIDITY
Can the retailer provide usable POS, inventory, markdown and waste data?
Can Departmint recommend early enough to matter?
Will staff actually act inside a normal shift?
ECONOMIC VALIDITY
Higher recovered margin?
Lower avoidable waste?
Acceptable markdown cost?
Acceptable staff effort?
Acceptable false-positive rate?
WHICH CATEGORY SHOULD WE START WITH?
ILLUSTRATIVE SCORING
Strawberries9/10
WASTE EXPOSURE — HIGH MARKDOWN FREQUENCY — HIGH DECISION WINDOW — HIGH
Blueberries8/10
WASTE EXPOSURE — MEDIUM VALUE DENSITY — HIGH DECISION WINDOW — MEDIUM
Avocados7/10
WASTE EXPOSURE — MEDIUM MARKDOWN FREQUENCY — MEDIUM DECISION WINDOW — MEDIUM
Tomatoes6/10
WASTE EXPOSURE — HIGH VALUE DENSITY — LOWER DECISION WINDOW — MEDIUM

The actual pilot category is selected from retailer data. Departmint does not assume strawberries are always the right starting point — this demo uses them because they make the timing problem legible.

11 — WHERE THE PRODUCT GOES

Prove the decision.
Then automate it.

STEPS 3–7 NOT BUILT TODAY
STEP 1 — TODAY
Structured observations + retailer data
STEP 2 — MVP
Decision model
STEP 3
Computer vision
STEP 4
Environmental signals
STEP 5
Condition intelligence
STEP 6
Automated recommendations
STEP 7
Multi-store optimisation
SHELF CAMERA — PRODUCE 04FUTURE VISION MODULE
4 POTENTIAL ANOMALIES
DETECTED — 142 TOMATOES
SHELF OCCUPANCY — 82%
CONDITION TREND — DECLINING
FRESHNESS PASSPORTLONG-TERM VISION
BATCHST-240812
ORIGINNETHERLANDS
ARRIVAL UAE10 AUG
DCDUBAI
STOREDUBAI MARINA
CONDITION HISTORYTRACKED
COMMERCIAL RISKMEDIUM
This dataset does not exist today. It is what enough deployments could eventually build.
GROUP VIEWFUTURE MULTI-STORE
DUBAI MARINAAED 1,840
ABU DHABI YASAED 1,120
RIYADH NORTHAED 1,460
JEDDAH CENTRALAED 780
DOHA WESTAED 640
INVENTORY AT RISK, BY STORE
ACTIVE INTERVENTIONS — 22
RECOVERED MARGIN — AED 6,340
12 — BUSINESS MODEL

Built as enterprise software.

01 — DESIGN PARTNER PILOT
One store, one category
A controlled workflow with terms agreed with the partner.
02 — STORE DEPLOYMENT
Annual subscription
Potentially priced by store, module and category.
03 — ENTERPRISE
Multi-store agreement
With integrations, analytics, implementation and support.
LAND AND EXPAND
One category
Multiple fresh categories
Multiple stores
Retail group
Other perishable departments
FRESHBAKERYMEATSEAFOODDAIRY
PRICING
Custom enterprise pricing.
Pilot terms agreed with selected design partners.
GCC FOOTPRINT — INTENDED
UAESAUDI ARABIAQATARKUWAITBAHRAINOMAN
First design partnerships are intended for GCC retailers, starting in the UAE and expanding into Saudi Arabia.

Want to test this on real fresh inventory?

We're looking for GCC retailers willing to test one category, one store and one measurable decision — against their own current process.

Become a design partnerTalk to Departmint
Every number, dashboard and outcome on this page is simulated. Departmint has no deployed retailers, no customer results and no validated accuracy.
Investor? View the investment thesis →