Decision Intelligence Engine
Continuously interprets unified SKU × Store × Day data and surfaces the highest-value actions — ranked by the value at stake, not the volume of alerts.
Diagnose · Predict · Prescribe · Automate · Monitor
Keystone is now KeyOne
The chain already knows.
From raw retail data to shelf-level action
KeyOne unifies your SAP and retailer scan data into one source of truth at SKU × Store × Day — then turns every signal into a tracked, closed-loop decision your field team executes and measures.
Signal A shelf, a scan, a rep’s photo — something changes.
Priority Ranked against every other open signal, by revenue.
Decision The engine — or a person — decides what to do.
Action Routed to the one person who can move it.
Outcome Measured back. Nothing closes itself.
Every day your chain makes thousands of decisions like this one. KeyOne makes sure they get made — and made well.
What sets KeyOne apart
Not dashboards. Decisions.
Most analytics tools stop at Diagnose. KeyOne carries every problem all the way through Prescribe, Automate, and Monitor — and proves the result.
Closed-Loop Tracking
Every signal becomes a tracked work item: detected automatically, assigned to a rep, executed in-store via Field, and measured. Nothing is raised and forgotten.
Work Graph · Decision lifecycle · Verified outcomes
KeyChat — Conversational AI
“Why did rate-of-sale drop at GreenMart in Gauteng?” Get a narrative answer with inline charts, grounded in your bound data. Every figure comes through a read-only tool over your own warehouse, with the query one click away.
Read-only tools · Drill-downs · Cited answers
One Source of Truth
SAP/ERP and retailer scan data read from the warehouse you already run, at SKU × Store × Day. Every team acts on the same number — and the query behind it is one click away.
Canonical vocabulary · Warehouse binding · Query behind every number
Field — In-Store Execution
Mobile-first Win List delivered to every rep. Beat plans, shelf audits, and Image Recognition ensure the boardroom decision reaches the shelf — and the result is measured.
Offline-capable · Ranked Win List · IR audit
Multi-Tenant & POPIA-Ready
Each principal is an isolated tenant, enforced with row-level security in the database. Built for South African data residency requirements from day one — not retrofitted.
Tenant isolation · Audit trail · POPIA s22
Commercial Cockpit
Decision hubs.
One commercial picture.
Every hub answers a specific commercial question with ranked decisions, not raw data tables.
Command Centre
“How is the business tracking against target?”
Sales Out
“How are we doing against last year?”
Price & Promo
“Are prices in tolerance and promotions reaching the shelf?”
Margin
“Where is gross margin leaking?”
Trade Investment
“Where did trade spend go, and what did it buy?”
Projections
“Where are we landing against this period?”
Availability
“Where are we out of stock despite being ranged?”
Distribution
“Where are we losing listings?”
Replenishment
“Is service level being met, and where is cover thin?”
Phantom Stock
“What is on the book and not selling, and what was written off?”
Listing Status
“What has the retailer done to the range — listings gained, lost or recoded, and discontinued or negative stock?”
Perfect Store
“Is the shelf right — facings, share, compliance?”
Field Execution
“Are reps covering their calls and completing them?”
1Every hub · one reconciled commercial picture
Not slideware
The actual product.
Screenshots, not renders.
Four more real screens from the same build — the cockpit, the field app, and the people running both.
Every decision, with its evidence
Value impact, entity evidence, and why there’s no "mark done" button — only the next read of the data closes a card.
The team behind the numbers
Rep roster and outlet coverage — open a rep to see the decisions and visits waiting on them.
The visit, not just the plan
KeyLink turns a route stop into a Win List — open actions and opportunity value, per store.
Every gap, ranked
Every gap the platform has detected in your own data, ranked by value impact — filterable by hub, severity, and status.
Platform architecture
Domain depth.
Platform flexibility.
KeyOne pairs deep FMCG decision logic with one service that reads the warehouse you already run — so you get domain depth without a second copy of your data.
Interfaces Where your team acts — browser, mobile, or conversation.
Commercial CockpitField AppKeyChat AIDocs & Admin
KeyOne Domain Pack FMCG decision logic, engines, and closed-loop tracking — KeyOne’s IP.
Decision EngineAvailabilityDistributionPrice & PromoMarginTrade InvestmentReplenishmentPerfect StoreField ExecutionProjectionsCommand CentreSales OutPhantom StockListing StatusWork GraphMaster Data AdoptionShelf-Image Intake
Core Service One Go service that reads the warehouse you already run — the foundation KeyOne runs on.
Canonical VocabularyWarehouse BindingPostgres · BigQuery · Snowflake · Databricks · AthenaDefinition CompilerScheduler & Run HistoryRow-Level Tenant IsolationNotifications & DigestsRead-Only KeyChat Tools
Why a domain pack matters
Rather than building a bespoke data lake, KeyOne binds a fixed canonical vocabulary — sixty named fields across products, outlets, sales, stock, shelf and visits — to the tables you already have. Every hub, rule and drill-down is written once against that vocabulary and compiled to your warehouse’s SQL.
What the core service provides
Tenant isolation with row-level security, a definition compiler for five warehouse dialects, a scheduler with a run history you can read, notifications and digests, and read-only tools for KeyChat — with the FMCG vocabulary layered on top.
The domain is deep. Decision hubs, a closed-loop work graph, and master data adopted from your own facts, built specifically for FMCG route-to-market — not a generic BI template.
The core is real. Row-level tenant isolation, a canonical vocabulary compiled to your warehouse, and read-only tools underneath every hub and every KeyChat answer — not a bespoke one-off build.
The market is the proving ground. Built for South African data residency and POPIA from day one, for principals who operate across multiple retailers.
How it works
Decide → Act in-store → Re-measure
KeyOne closes the loop between a boardroom insight and a verified shelf outcome — every single time.
01
Diagnose
What is happening, and why?
02
Predict
What will happen next?
03
Prescribe
What should we do about it?
04
Automate
Routed to whoever can act — nobody forwards it.
05
Monitor Differentiator
Did the action work?
Most analytics toolsStops at Diagnose
KeyOneCarries through to Monitor
Most analytics tools stop at Diagnose. KeyOne's differentiator is carrying the same problem all the way through Prescribe, Automate, and Monitor — and proving the revenue result with closed-loop measurement.
Built for South African FMCG principals — the people who own distribution, field, and trade.
Closed-loop
Signal → decision → outcome
SKU × Store × Day
Granularity of truth
POPIA
Data residency by design
Peer-compared
Outcomes measured against untreated stores
Designed for principals who operate across multiple retailers and need one trusted cockpit — not seven separate portals. Every number in KeyOne is traceable to source, every decision is tracked to outcome.
— Built for SA FMCG principals · Cognizance Processing (Pty) Ltd
Turn your retail data
into decisions that ship.
Replace fragmented spreadsheets and retailer portals with one trusted, closed-loop platform — from commercial cockpit to rep Win List.
Built for South African retail · POPIA-compliant by design · No lock-in