From everyday receipts
to rich product identities
Receipts bring purchase data into Genie Vault. Our engine combines that data with product knowledge to build detailed descriptions, assign categories and identify product features.
Catalogues · databases · market information
Everyday receipts
A growing stream of purchases, shops and product descriptions.
Genie Vault
Purchase records brought together for processing.
Genie’s product
intelligence engine
Read the purchase
Extract the available product, shop, price and quantity details.
Build the description
Combine receipt details with available product knowledge.
Categorise and describe
Assign a category and identify the product’s distinguishing features.
Structured product knowledge
Product identity
- Recognised category
- Brand and family
- Variant and features
- Specific pack
Connected to the original purchase records.
One consistent view of what was bought
Detailed product identities enable comparisons, alternatives and clusters across brands and retailers.
Refined through training and testing.
Evaluation helps improve how the engine interprets and describes products.
Every product.
Its identity, features and connections.
Once a product has been categorised, Product Unifier builds a more detailed identity — connecting purchase records, distinguishing variants and revealing comparable products across brands.
From category to product identity
After category recognition, each level adds detail about what someone bought and how it relates to other products.
Type, range, line and item form the product structure within the recognised category.
A complete product identity
Orchard Free strawberry yoghurt
150 g pot
- Product type
- Fruit yoghurt
- Format
- Spoonable
- Flavour
- Strawberry
- Texture
- Smooth
- Fat
- 1.5%
- Lactose-free
- Confirmed
The same identity can link descriptions from several shops. Each purchase keeps its own date, price and original record.
Recognise repeat purchases
Different retailer descriptions can resolve to the same product.
Find meaningful alternatives
Compare the full combination of features across brands and pack sizes.
Understand buying patterns
Explore purchases at the level of a product, family or shared feature cluster.
One category. Many precise identities.
Within the recognised category, the product tree adds progressively finer detail. Each variant combines its own features with those of its family.
Select an item to follow its path and see its complete set of features.
Similar products have different relationships
A shared identity joins descriptions of the same item. Shared features connect separate items without merging them.
One catalogue item
D · Orchard Free strawberry 150 g
Two retailer descriptions, one product identity. Two purchases remain two purchases.
| Product | Flavour | Texture | Fat | Lactose-free | Pack | Relationship to D |
|---|
“Strawberry” is only one feature. Meadow Free can be a closer alternative to Orchard Free than Orchard Classic is, even though the two Orchard products share a brand.
Shared features bring products closer
The tree follows product families. Clusters cut across those branches to reveal a different kind of connection.
A precise group, across brands and packs
Combine features to select a product cluster. The group can span several families while retaining every individual product identity.
More relevant alternatives
Search across brands using the features that matter to the buyer.
More useful comparisons
Keep pack sizes visible and compare unit prices with their shop and date.
Deeper purchase analysis
Build purchase-based cohorts from these product groups. A purchase shows buying behaviour, not a medical condition.