PROJECT — RESTRICTED

Redacted.

AI orchestration completely embedded across the operations of a UK food & beverage manufacturer. Cannot show. Has changed how they operate.

One operator. Multiple connected surfaces — supply chain, demand planning, sales, market intelligence, commodity & FX exposure, financial reporting. Designed end-to-end. Built solo. Live with enterprise users across the C-suite and operating functions.

What follows is the application, with customer name, product names, and counterparty names blacked out. The numbers stay, because the numbers are the point — they show the scale of what one platform is doing across one organisation.

Customer name, product names, counterparty names, and the platform’s own brand are redacted under NDA. The work is real. The screens are live.

Operational overview dashboard — The overview surface. Revenue, customer count, order volume, and active alerts sit at the top, with weather-and-supply risk signals across the commodity book underneath. The alert feed runs live.

Demand planning dashboard — Demand planning. Forecast demand, coverage gaps, reorder value, active SKUs. Top risk categories and replenishment priorities are surfaced for the planning team — the agent does the prioritisation, the human approves.

Supply chain stockouts and category heatmap — Supply chain. Critical stockouts, warning stockouts, overstock, open POs. The eight-year category volume heatmap below grounds every alert in long-running seasonality.

Product margin trend dashboard — Margin intelligence. SKU-level margin trends, twelve-month rolling averages, and an Explain This affordance on every chart — the agent narrates why a margin is compressing or expanding, in plain English, grounded in the actual data.

Sales / customer book — The customer book. Channel, region, status, last-activity, lifetime value, current state. Revenue by sales rep and revenue by channel sit underneath. Sales reps work the same surface as planners and finance — one platform.

Market news intelligence feed — External signals. Market news scraped, classified, and tagged against the operation’s actual exposure — energy, supply, demand, regulation. Each item is a potential trigger; the planning surfaces respond to what gets confirmed.

Commodity and FX corridor view — FX exposure. Twelve-week rolling corridors per pair, current position, hold/move recommendation. Tens of millions in currency exposure routed against actual purchase orders, not generic hedging rules.

Reports / financial performance — Reporting. Revenue trend across three years, customer counts, and an Ask-for-any-report prompt where the C-suite types what they want and the agent compiles it from the underlying data.