The Advise Engine
Everything Advise does runs on one engine. It takes the messiest data in consumer goods, turns it into one trusted view, and computes every number your teams and your AI will ever show a buyer.
102 stores sold it a month ago. 40 still do.
62 dropped the 250ml can at the week 33 planogram reset. Store 41 is one of them.
Inside Store 41, the problem has an address.
Aisle 7, cold drinks. Bay 3, shelf 3.
Four empty facings. The tag is still there.
Kestrel Energy 250ml left the range at the week 33 reset. Nobody told the shelf.
Every tag is a barcode. Every barcode is a record.
Follow the bars and you are inside the POS data.
Every shelf is a row. Most teams never see this one.
Week 33, Store 041, Kestrel Energy 250ml. Units 0. Ranged: no.
The AI never runs a query. It never invents a number.
Intelligence chooses three analyses from a governed Function Registry. Core computes every figure.
A decision you can take into Thursday’s buyer meeting.
Reinstate the 250ml in the 62 stores delisted in week 33. The evidence comes with it.
Make better decisions, faster.
From category to product to store, and back again. Every Walmart store in Dallas–Fort Worth is lit again.
What you just zoomed through
Every store, every shelf and every scan is a row. Most teams never see the row that explains their problem. The engine finds it, connects it to everything else, and hands it up to the analytics and the AI.
1. Ingest and code every feed
Advise takes your data in the formats it actually arrives in, then codes it once:
- Retailer EPOS / POS: store-level and week-level, from retailer portals and files
- Household panel: shopper behaviour, penetration and switching
- Loyalty data: baskets, visits and customer segments where the retailer provides them
- Price and promotion: captured daily from retailer websites and matched to your sales
- Orders and financials: EDI orders and ERP extracts
- Field data: audits, photos and compliance from the Sales Assist app
Products, stores, weeks and retailers are mapped to one product master and one calendar, so "Kestrel Energy 250ml" means the same thing in every source. You control the product attribution in one place, and a rules-based auto coder codes new and unmatched SKUs, so coverage keeps rising as new lines arrive.
2. One View
The result is a single harmonised model: your category, your brands, your retailers and your competitors on one version of the truth. Metrics are built to match how each retailer measures you, so the number in your deck is the number in the buyer's.
3. The deterministic analytics engine
Dashboards, trackers, elasticities, promo ROI, distribution gaps and store rankings are computed by code, the same way every time. Nothing is estimated by a language model.
4. The governed AI layer
Intelligence and Autonomous reach your data only through our Function Registry: a governed library of analyses and functions available to the AI. The AI has access to a toolbox of tools.
- The AI chooses which analyses to run and in what order.
- Core computes them on your data.
- The AI explains the result in plain language, with every figure traceable to its analysis.
The AI never runs a query and never invents a number.
Where it runs
- Model: Microsoft Azure OpenAI, served within your data region. It is secure and cost-efficient, and it never learns from your data.
- Advise Cloud: hosted in Advise's Azure tenancy, a dedicated environment per client, in your chosen region.
- Advise Private: installed in your own cloud tenancy, where you own the data, the keys and the install. Talk to us
- Access: role-based. People see only the data they are allowed to see, and so do the AI layers working on their behalf.
Years in production
The engine has run in production since 2020 for brand owners including Birds Eye, Coca-Cola HBC, Tirlán and O'Brien Fine Foods.
The same engine, for retailers
Advise Retail uses the same engine for retailers' own analytics and for providing data to their suppliers. See Advise Retail
Related
Advise in 90 seconds
One source of truth for your retail data, more than 50 dashboards, and AI that explains them. The quickest way to see what Advise does.
WatchHow Britvic Ireland turned strategy into measurable results
Britvic Ireland used Advise to sharpen commercial priorities, improve in-store execution and make trade spend more efficient across sales and category teams.
Read the case studyFinance doesn't need more spreadsheets. It needs one version of the truth.
CPG finance teams track margin, trade spend and promo ROI across retailers on fragmented data. How one view aligns finance with category, RGM and sales.
Read more
1:33Make better decisions, faster.
See Advise in its full glory in a 30-minute demo, or explore the sandbox first.
