The AI Operating System
for Retail Execution
Not an audit app. The system of record for what happens on the shelf — capture, detection, orchestration and analytics as one loop.
Most retail execution breaks between the store visit and the decision.
Brands invest heavily in shelf space, promotions and field teams, yet most only see what happened on the shelf weeks later. By then, the sales are already lost.
No live view of the shelf
Audits and photos pile up, but no one can say which stores are on plan today and which are losing sales.
Issues found, never fixed
Gaps get reported, but without a clear owner, deadline and proof of fix, they stay on the shelf.
No single source of truth
Audits, tasks, planograms and sales data sit in separate tools, so leadership never sees the full picture.
Operating System for Merchandising Execution
Not an audit app. The system of record for what happens on the shelf — capture, detection, orchestration and analytics as one loop.
One photo carries the whole context
Every capture is stamped with store, aisle, planogram version, scanned SKUs and coordinates before it leaves the device. Nothing depends on the rep remembering to fill a form.
The image returns a verdict, not a file
Mechy checks every photo against the planogram assigned to that store on that date and tells the rep exactly what to fix, while they are still in the aisle.
Every gap leaves with an owner
Each finding becomes a task tied to the exact SKU, shelf and position, with an owner and a deadline. A task closes only when a verifying photo passes the same check.
The loop pays for the next loop
Each cycle adds labelled shelf data. Models sharpen on your categories, forecasts tighten, and schedules shift toward the stores where a visit is worth the most.
Other Key Features
One intelligence layer for every critical part of retail execution.
Agents that work the shelf while your team works the store
Each Mechy agent watches one recurring job, runs it on schedule against live shelf data, and hands your team a finished answer or an assigned task.
Monday shelf report
Last week’s compliance by region, category and rep, written up before the first call.
Promotion check
Confirms every promoted display went up on time and at the agreed location.
Out-of-stock recovery
Spots empty facings, checks order history and raises a suggested order.
Planogram reset
Tracks a new planogram from release to verified install across every store.
What changes in the first two quarters
Planogram compliance across 1,240 stores
Replaced a monthly photo audit with continuous scoring. Field visits stayed flat.
READ THE CASE STUDY →Average time from detection to verified fix
Down from nine days of email chasing between HQ and regional leads.
READ THE CASE STUDY →Out-of-stock hours on promoted lines
Availability gaps routed to store teams the same hour they were detected.
READ THE CASE STUDY →“We stopped arguing about whether the display went up. The photo, the score and the fix are in one record, and the category team reads it the same way we do.”
“Our reps were spending the last hour of every visit filling forms. Now the capture is the report, and the hour goes back into the store.”
“The value showed up in the second quarter, when the model started telling us which stores to visit instead of us guessing.”
Questions buyers ask first
If yours is not here, bring it to the demo.
Baseline models ship at 94–96% on packaged goods. Accuracy is re-measured on your own labelled set during onboarding, and the model is fine-tuned per category before rollout.
A pilot across 20–50 stores runs in three weeks: planogram import, user and territory setup, one training session for field teams. National rollout typically follows within a quarter.
Most teams retire their audit tool and keep their SFA. Merchy can run as the capture and detection layer behind an existing app through the API, or as the full field client.
Regional cloud tenancy with SSO, role-based access and audit logs. Images and scores export to your warehouse on a schedule, so the analytics stay yours.
Captures, checklists and scans queue on the device with their GPS and timestamp intact, then sync in order when connectivity returns. Nothing is re-entered.
Every store, every shelf, executed to plan.
A 30-minute session on your categories, your planograms and your stores.
Four stages, one record. Here is what happens between the photo and the fix.
Most tools stop at one stage and hand the rest to email. Merchy holds the whole cycle, so every artefact — image, score, task, outcome — points back to the same store visit.
Field teams photograph the shelf
Barcode, GPS and offline built in. The rep opens the visit, shoots the bay, and the app attaches the store, the planogram version in force that day, the scanned SKUs and the coordinates.
AI scores every display against the planogram
Recognition returns facings, share of shelf, adjacency, price card presence and out-of-stocks in under three seconds. Confidence below threshold routes to human review rather than guessing.
Gaps become assigned tasks, tracked to close
You define the rules once: which detections create work, who owns them by store and region, how long they have, and who hears about it when the clock runs out. Closure needs a verifying photo.
Compliance, sales and forecast data feed the next cycle
Every closed loop is training data. Detection sharpens on your packs, forecasting learns which gaps cost the most, and the next route plan sends reps where the return is highest.
SEE THE FULL FEATURE SET →Shelf data in, decisions out
Merchy reads from the systems that define the plan and writes to the systems that act on it, on a daily schedule or in real time through the API.
Underneath the loop
The parts a platform team asks about in week one.
See the loop run on your own stores
One intelligence layer for every critical part of retail execution.
Twelve capabilities, three groups. Each one writes to the same execution record, which is why the whole is worth more than the parts.
Barcode Scanner
Confirm product-level execution in seconds.
Scans resolve to your SKU master, so a listing gap is recorded against the exact article number rather than a free-text note.
GPS / Location
Connect proof of work to every store.
Coordinates and timestamp are written at capture and cannot be altered afterwards, which settles most disputes before they start.
Offline Ready
Keep work moving beyond the signal.
Visits complete in full without connectivity. The queue syncs in capture order when the device is back on the network.
iOS & Android
Empower every field team on any device.
The same flow on company handsets and BYOD, including mid-range Android, where most field programmes actually live.
Communication
Keep teams and operations in sync.
Threads attach to the store and the task, not to a person's inbox, so context survives when territories change hands.
Multi-Language
Roll out better retail workflows globally.
Interface, checklists and task text localise per user, while the underlying record stays in one canonical language for reporting.
Image Detection
Understand every shelf from a single photo.
Facings, share of shelf, adjacency, price cards and out-of-stocks, returned with a confidence value on each finding.
Planograms
Keep every store aligned to the shelf plan.
Plans are versioned by store, channel and date, so a score is always judged against the layout that was actually in force.
Smart Alerts
Surface the moments that need attention.
Thresholds are set per category and region so alerting stays useful rather than becoming another channel people mute.
Task Tracking
Assign, verify, and close actions fast.
Ownership, SLA and escalation are rules, not reminders. A task closes only when a verifying photo passes the same model.
Orders
Spot availability gaps before they escalate.
Detected out-of-stocks can raise a suggested order against the store's own history instead of waiting for the next call cycle.
Forecasting
Turn field intelligence into better decisions.
Compliance history joins sell-out data, so planning sees which shelf conditions moved volume and which visits were worth making.