Home/AutRi
Recent Innovations · Agentic Planogram Compliance · Grocery

AutRi.

AutRi captures shelves; GMetri reads them against the planogram; the Agentic Command Centre routes the fix to PulsePoint — and the next AutRi scan closes the ticket.

Pilot · Phase 1 · FreshPik Powai with Fynd Nucleus Building · agentic loop end-to-end Roadmap · grocery-estate rollout
Today · running at FreshPik Powai · & the agentic-loop targets
Shelf compliance · today
95%+
measured at FreshPik Powai pilot
SKU recognition · today
98%
per-SKU bounding boxes
Coverage per scan run
20–30K sqft
single-store sweep
Verified-closure target
< 24 hr
next scan confirms the fix held
AutRi autonomous shelf-scanning robot navigating a grocery aisle · per-fixture image capture with on-board AI bounding-box detection · light beams visualise the scan path
AutRi · autonomous shelf-scanning robot · navigates the grocery aisle and captures every fixture per scheduled session. 20–30K sqft per run.
01 · Status

Today: a deployed scanner. Next: a closed loop.

Phase 1 with Fynd Nucleus is running at FreshPik Powai — AutRi executes scheduled scans, draws bounding boxes around recognised SKUs, and produces brand-occupancy and facing heatmaps. A human still has to log in, interpret the heatmap, and tell someone to fix what's off. The agentic flow removes the human from the middle of that loop.

Today · FreshPik Powai
Pilot

Standalone analytics tool.

  • Scheduled scan sessions captured per zone and fixture · 20–30K sqft per run
  • Internal CV models draw SKU bounding boxes against AutRi's own SKU master · 98% recognition
  • Output: brand-by-shelf-level + facing-count heatmaps in AutRi's dashboard
  • Human loop: a Cluster Manager logs in, interprets the heatmap, manually raises the fix request
In the agentic flow
Building

High-fidelity producer + loop closer.

  • No more human interpretation · tagged shelf captures flow directly into GMetri via API
  • Context-aware scanning · schedule dynamically tied to restocking windows + the promotion calendar
  • The closer of the loop · a PulsePoint ticket marked "Done" stays open until AutRi's next scan of that fixture confirms compliance
  • Trust comes from verification, not from the staff member's word
AutRi Sessions dashboard at FreshPik Powai · 42 total sessions, 11 active, 17 completed, 11 failed · scheduled scan list with per-session status
Sessions surface · scheduled scans across the store, with per-session status.
AutRi session detail · scan tasks broken down by 50 grocery categories at a pilot store, 98 of 100 done with category-level status
Session-detail surface · 100 scan tasks broken down by 50 grocery categories.
AutRi analytics output · Brand by Shelf Level stacked bar + Product Distribution Across Shelf Levels facing-count heatmap with intensity legend
Analytics output · Brand-by-shelf-level + facing-count heatmap. Useful, but the human still has to read it and decide what to do.
02 · The 6 compliance dimensions

What "compliant" means on a grocery shelf.

Planogram compliance means the live shelf matches the plan in every store, all day, every day — across six dimensions. Each one is what the AutRi scan reads, what GMetri interprets, and what the Agentic Command Centre decides on.

DimensionWhat the system is checkingExample non-compliance
Product presenceIs the SKU on the shelf at allEmpty slot for a top-velocity SKU
Product placementIs the right SKU on the right fixture and shelf levelWrong brand of cooking oil on the premium shelf
Facing countAre the expected number of facings present2 facings instead of the planned 4 for a promotional SKU
Price tag accuracyDoes the price tag match the current price masterOld price tag on a live promotional SKU
Promotional executionAre promotional end caps and POS displays set upMissing end cap for the week's featured SKU
Shelf cleanliness & orderIs the shelf set up as per the visual standardProducts toppled, stacked incorrectly, or damaged
AutRi high-density grocery shelf capture with AI bounding boxes detecting individual SKUs across the Global Flavours wall
Per-SKU detection on a high-density grocery wall. Every box is a candidate for one of the six checks above.
AutRi bounding-box detection across a cooking-sauces shelf — Sriracha, ketchup, sweet chilli, hot sauces with per-SKU labels and brand attribution
Per-brand attribution on a cooking-sauces fixture. Brand × shelf level × facing count are computed off the same capture.
03 · Architecture · 3 layers + 1 master data plane

One shared master data plane. Three layers above it.

Every layer exists to serve the planogram-compliance use case — no more, no less. The Data layer sees the shelf, the Intelligence layer judges it, the Action layer fixes it, and the Data layer re-sees it to close the loop.

Layer 1 · Data
Live

AutRi robot · Image Repository

Produces the ground truth of the shelf, tagged to fixture, zone, and timestamp. Per-scan-session uploads of high-resolution captures with bounding boxes.

Layer 2 · Intelligence
Building

GMetri · Agentic Command Centre

Turns ground truth into a compliance judgement, dimension by dimension, then decides what to do about it — triaged by SKU velocity and severity.

Layer 3 · Action
Building

PulsePoint Admin · PulsePoint App

Turns the decision into the specific act that restores compliance, with the right person, the right SOP, and an evidence requirement.

Master data plane
Building

SKU · Planogram · Store · Roles

The single source of truth for what compliance means on every fixture, in every store, for every role. Hourly sync; stale planograms block ticketing.

Architecture diagram · Master Data feeds Action and Intelligence; AutRi Robot + Image Repository under Data; GMetri + Agentic Command Centre under Intelligence; PulsePoint Admin + PulsePoint App under Action; verified-by-next-scan loop closes the cycle
Source diagram · 21-Apr-2026 brief

The dotted "verified by next scan" arrow from Action back to Data is the contract that makes this self-verifying — without it, this is just another ticket queue.

04 · The 5-step loop

Scan. Read. Decide. Act. Verify.

A ticket is closed only when the next AutRi scan confirms compliance on that fixture. Staff evidence alone is not closure. This is what makes the loop self-verifying — and what makes the compliance number trustworthy.

Step 1 · Scan
Live

AutRi

Robot runs scheduled sessions; captures shelf images per fixture and zone.

Output: tagged shelf images in the image repository.
Step 2 · Read
Building

GMetri

Reads the shelf against the active planogram version; detects presence, placement, facings, price, promotions.

Output: discrepancy list per fixture with evidence + confidence.
Step 3 · Decide
Building

Agentic Command Centre

Triages each discrepancy by compliance dimension, severity, and SKU velocity; decides the action.

Output: triaged compliance alert with SOP mapping and route.
Step 4 · Act
Building

PulsePoint

Auto-creates a ticket with the shelf image, expected planogram view, and SOP; routes to the right role.

Output: ticket on the right person's PulsePoint app.
Step 5 · Verify
Building

AutRi + PulsePoint

Staff evidence moves the ticket to pending verified. The next AutRi scan closes the ticket — or escalates.

Output: compliance confirmed, ticket closed, audit trail updated.
5-step loop diagram · SCAN → READ → DECIDE → ACT → VERIFY with a re-scan loop arc back to step 1
End-to-end sequence diagram · Store Ops scheduler triggers AutRi Robot; tagged images upload to Image Repository; GMetri reads against planogram; discrepancies flow to Agentic Command Centre; tickets created in PulsePoint; staff execute SOP and upload evidence; next AutRi scan re-reads the fixture and closes or escalates
End-to-end sequence · 17 numbered steps

The exact sequence the system runs, from a scheduled scan to a verified closure. Every step has an owner and a handoff — no manual bridges.

05 · Per-dimension handling

Six dimensions. Six routes. Six SOPs.

Each compliance dimension has a specific detection path, a specific severity decision, and a specific role it routes to. Velocity weighting and the active-promotion window override the default severity.

DimensionDetected by GMetriDecided by Command CentreActed on in PulsePoint
Product presence Empty or non-matching slot Weights by SKU velocity · flags Critical for top-velocity SKUs Restock SOP → Department Manager
Product placement Wrong SKU on fixture Flags High Planogram reset SOP → Department Manager
Facing count Facings counted vs planogram Flags Medium · High if SKU is promotional Facing correction SOP → CSA
Price tag accuracy Reads tag text · compares to price master Flags Critical for live promotional SKUs · High otherwise Price correction SOP → Store Manager
Promotional execution Missing or incomplete end cap vs planogram Flags Critical during the promotion window End cap setup SOP → Visual Merchandiser
Shelf cleanliness & order Toppled or out-of-order state Flags Medium Shelf reset SOP → CSA
06 · Plan · 4 stages

Four stages. Each stage is a standalone unlock.

Value starts flowing before the full loop is live. Stages run in sequence, with Stages 2 and 3 overlapping through the middle of the build.

Stage 1 · Foundation
Building

Shared data model.

Align fixture, zone, SKU, and store IDs across AutRi, GMetri, and PulsePoint. Load grocery SKU catalogue and planogram library into GMetri. Ingest the promotion calendar.

Exit: data model signed off by all three teams; planogram live in GMetri; promotion calendar ingested.
Stage 2 · Intelligence
Building

GMetri tuned for grocery.

Tune GMetri across all six compliance dimensions. Write grocery triage rules in the Agentic Command Centre — dimension, severity, velocity, promotion-window — and the SOP mapping.

Exit: GMetri reliable at shelf level; triage rules approved by grocery ops leadership.
Stage 3 · Action in PulsePoint
Building

Auto-ticketing end-to-end.

Build the intake from Command Centre into PulsePoint. Configure grocery SOPs, evidence requirements, and the escalation matrix from Store Manager → Cluster Manager → State Head.

Exit: auto-ticketing live end-to-end; SOPs configured; escalation matrix live; audit trail captured.
Stage 4 · Pilot & scale
Roadmap

Two pilot stores → chain.

Run the end-to-end loop in two pilot grocery stores. Measure loop health weekly. Scale to the rest of the grocery estate after four consecutive weeks of steady KPIs.

Exit: pilot KPIs steady for four consecutive weeks · sign-off for chain-wide rollout.
07 · Design choices for grocery

Three choices that make this work for grocery.

Grocery is not fashion. The design choices below earn trust with store staff from day one and keep the inbox focused on what hits revenue.

Choice 1

Velocity-weighted severity.

An empty slot for a top-velocity SKU is not the same as a facing drift on a slow mover. The Command Centre weights triage by SKU velocity, not raw count of violations. Keeps the Store Manager's PulsePoint inbox focused on what hits revenue and customer experience.

Choice 2

Promotional execution is its own lane.

Promotions drive a disproportionate share of grocery sales and customer trust. Price mismatches and missing end caps during an active promotion window are flagged Critical and go as push, regardless of the underlying compliance dimension. This lane has its own SLA.

Choice 3

Scans timed to restocking.

Compliance signal is only meaningful after restocking has happened. AutRi scans are scheduled immediately after the primary restocking windows so alerts reflect post-restock reality. This single choice removes a large class of false alerts and earns trust with store staff from day one.

08 · Outcome targets

Compliance rate is the headline. The loop is the product.

Nine targets, every one measurable. They are aspirational until the loop is live in pilot — then they become the contract.

OutcomeTargetWhy it matters
Planogram compliance rate · grocery estate · velocity-weighted · all 6 dimensions≥ 95%Direct measure of shelf health
Time to ticket · scan → PulsePoint ticket created< 10 minTells us the intelligence layer is responsive
Time to first action · ticket created → staff acknowledgement< 30 min
Critical & High		Tells us the action layer is reaching the right person
Time to verified closure · ticket created → next scan confirming compliance< 24 hrTells us the loop is actually closing, not just marked closed
Share of tickets auto-created vs manually raised≥ 90%Tells us the system, not the human, is driving the loop
Re-open rate after staff marked closed< 5%Tells us evidence is not being gamed
Manual shelf-audit effort · per store per week−70%Direct operational saving for Cluster Managers
Out-of-stock rate · top-velocity SKUs−40%Business impact on revenue
Promotional price & end-cap errors caught inside the hour≥ 95%Customer experience and trust
09 · Risks & how we design against them

Five named risks. Five named mitigations.

The loop only works if the staff trust it, the data stays fresh, and a robot outage doesn't stop the store.

RiskHow we design against it
AI noise on dense shelves · visually similar SKUsConfidence threshold inside the Agentic Command Centre. Low-confidence items queue for a quick review before becoming a ticket. Continuous calibration on misses.
Staff alert fatigueStrict severity tiers. Critical and High go as push. Medium and Low batch into one daily task per zone. One-pager per SOP, no ambiguity.
Planogram or promotion-calendar driftHourly sync of planogram library and promotion calendar. Block ticketing on any fixture whose planogram version is stale.
Scan-coverage gaps when the robot is offlineGraceful fallback to a manual audit checklist in PulsePoint for affected zones. Auto-resumes once the robot is back online.
Role-mapping errors routing tickets to the wrong personMaster data plane owns the role map, with daily reconciliation against HRMS. Incorrect routes get flagged in the Command Centre.