Use case · FMCG / CPG

Retail Execution & Shelf Reality OS

An agentic layer that turns store visits, photos, and audits into ranked, store-specific fixes — so field teams act on shelf reality, not paperwork.

Shelf ReaderVoid PrioritizerPromo Compliance CheckerAction DrafterShare-of-Shelf Estimator
The friction

Where value leaks today.

Perfect store programs die in the gap between what head office plans and what actually sits on the shelf. Field reps walk stores armed with planograms and checklists, photograph the aisle, log out-of-stocks, and file an audit — and then that data lands in a reporting tool where it ages. The void detected on Tuesday gets a corrective action on the following Monday's call, by which time the display is gone and the sale is lost.

A retail-execution app digitizes the visit. It does not decide which of forty flagged issues a rep should fix in the eleven minutes they have in the store, doesn't draft the case to the store manager, and doesn't chase whether the fix held. The intelligence that converts a shelf photo into a ranked, store-specific action lives in the rep's head and evaporates when they move territories.

Value leaks three ways: distribution voids that nobody closes fast enough, compliance failures on promotions you already paid for, and field hours burned on data entry instead of selling. Another app to fill in faster doesn't recover those hours — it adds to them. What's needed is a system that reads shelf reality and hands each rep a short, ranked list of fixes worth making before they walk out the door.

How it runs

One governed flow — agents act, you approve what matters.

Reps spend their store minutes closing the gaps that recover real volume, and head office sees whether each fix held instead of trusting that the audit equals the outcome.

kitsune os · retail-execution-shelf-reality-osrunning
01
Read the shelf
images, audits, out-of-stocks
Agent
02
Score gaps
voids, share, promo compliance
Agent
03
Rank fixes
by recoverable value per store
Agent
04
Draft the case
order, reset, manager ask
Agent
05 · gate
Rep acts in store
approves and executes the fix
Human
06
Verify & follow up
confirm the fix held
Agent
Agent — autonomousHuman gate — your approval
What the OS runs

One operating layer — eight governed jobs.

Each is a governed agent inside the same system, sharing context — not eight tools you stitch together.

01

Shelf Reader

Turns aisle photos into structured facings, voids, and price reads, so the visit produces a measured shelf state instead of a checklist someone has to re-key.

02

Void Prioritizer

Ranks distribution gaps by recoverable volume and store, so a rep fixes the three voids that matter rather than logging forty that don't.

03

Promo Compliance Checker

Compares what's on shelf against the agreed feature, display, and price for each event, flagging paid promotions that never went live.

04

Action Drafter

Writes the store-specific play for each visit — the order to suggest, the reset to make, the case to put to the manager — before the rep walks the aisle.

05

Share-of-Shelf Estimator

Measures your facings against the category and competitors from the same image, so reps argue for space with evidence, not assertion.

06

Route Sequencer

Orders each rep's day around where recoverable value sits, weighting high-opportunity stores over rote coverage of the territory.

07

Fix Verifier

Checks whether last visit's correction held on the next pass, so chronic problem stores surface instead of quietly reopening.

08

Field Coach

Spots where a rep or territory consistently misses voids or compliance and routes targeted guidance rather than blanket retraining.

Governed by design

Autonomy you can trust — because the control is built in.

The system acts on its own and every action stays legible, bounded, and reversible. You don't choose between speed and control; the control is what makes the speed safe.

Legible

See what was done, what was declined, and exactly what's waiting on you — nothing happens in a black box.

Bounded

Agents act only within the rules you set. Anything material or irreversible stops at a human gate.

Reversible

Every action is logged and undoable. A wrong turn is caught and rolled back, not discovered weeks later.

Owned

One operating system you own — not a swarm of rented agents you have to police. Built, run, accountable.

The outcome

Reps spend their store minutes closing the gaps that recover real volume, and head office sees whether each fix held instead of trusting that the audit equals the outcome.

In-visitFixes ranked at the shelf
1 OSAudit to action joined
VerifiedFixes checked next pass
OwnedNot rented per rep
Questions, answered

What you're actually getting.

Is this a product or a build?

A build. Kitsune forges a retail-execution operating layer around your planograms, promotion calendar, and field routes — owned by you, not a per-rep license on a generic visit app.

What stays in my control?

The rep does. Agents read the shelf, rank the gaps, and draft the case; the rep decides what to act on and makes the fix in the store. Nothing is changed at shelf without a person executing it.

How is this different from a retail-execution app?

An app records the visit and files it. This reads the shelf, ranks the fixes by recoverable value, drafts the play, and verifies the correction held — turning a logging chore into a short list of moves worth making.

Does it work from store photos alone?

Photos are the primary input. The Shelf Reader extracts facings, voids, and prices from aisle images and combines them with audit logs and the promotion calendar, so a single visit yields a measured shelf state.

How does it know a fix actually stuck?

The Fix Verifier checks each prior correction on the next visit, so stores that quietly revert get surfaced and re-prioritized instead of being marked done and forgotten.

Bring us the bottleneck.

We'll forge the operating layer around your friction — built, owned, and running.