CONSUMER PACKAGED GOODS

On-shelf availability control tower for a consumer-goods manufacturer

Every night a consumer-goods company receives retailer sales and inventory feeds, its own stock and orders, and the promotion calendar. Wexa opens a case for every empty shelf, finds the cause, and proposes the fix for the key account manager to approve.

Download this case as PDF Book a demo

tokens per case vs the same agents without Wexa (59,269 → 16,559)
tokens when orchestrated across clouds over A2A
model cost per case
graph nodes and 8,300 relationships

CLIENT: Multi-brand consumer-goods manufacturer (54 products, 3 retail accounts, 60 stores)
REGION: US
WHEN: 2026
ENGAGEMENT: Solution build
WEXA COMPONENTS: Wexa · context graph · 5-agent process flow · A2A across clouds · approvals · data tables

01Case 08

The challenge

Out-of-stocks, phantom inventory and missed promotions found too late
Signals spread across retailer feeds, DC stock, POs and promotions
Agents running in different clouds without a common control plane

What Wexa built

Five agents per case: exception triage, demand sensing, supply planning, retail execution and account action
Agents traverse the graph from the shelf to stores, products, DCs, orders and promotions
Typed outputs become proposed actions with master-data identities, never model text
Replenishment orders, field-rep tasks and briefs written only after the account manager approves
Agents deployed on Google Cloud and configured for AWS and Azure; Wexa as the control plane

02Case 08

How it runs on Wexa

FigureHow this use case runs on Wexa: client systems connected, ontology discovered, context graph in the Wexa memory layer, agents and process flows, and governance. Highlighted components are central to this use case.

03Case 08

Results

Measure	Before / without Wexa	With Wexa
Analysts working shelf exceptions	8 analysts	3 analysts
Annual cost of the process	$720K	$270K
Lost sales recovered per year (60 stores)	–	$1.2M
Time to implement the AI solution	6 months	4 weeks
AI build cost (one-time)	$480K	$30K
AI operating cost per year (models, infrastructure, upkeep, platform)	$312K	$164K
Total AI cost in year 1	$792K	$194K

With Wexa

net annual value to the client after Wexa's own operating cost

faster to implement

lower AI operating cost

less AI spend in year 1

Decisions agreed between the two arms in 16 of 20 cases; 4 of 4 when orchestrated across clouds
An approved replenishment order was written to the system of record and read back intact
Demand drawn from a real public retail sales dataset