AI Assistant (Leat MCP)
July 20
→
Product recommendations
Use customer interests, the current basket, product availability, and margin to choose what is most worthwhile to recommend. Products can be suggested without an incentive, helping the business increase sales without automatically giving value away.
Recommendations begin with affinity scores for the products, categories, and brands most relevant to each customer. These scores can reflect purchases, browsing activity where available, redemptions, and preferences the customer has shared. Leat can recommend familiar products or identify a related category that interests the customer even when they have not purchased from it before.
A useful recommendation reflects what is happening now. Leat can consider what complements the current basket, what the customer usually adds but has not selected, what is available at the location, and which channel they are using. Time and order context can also matter, such as suggesting a breakfast add-on in the morning, the correct size for a previous purchase, or an additional item that completes an order.
Each recommendation becomes a candidate scored according to its expected uplift, margin, availability, and cost. A relevant product that needs no discount carries no incentive cost and may rank above a discounted alternative. When two items are expected to perform similarly, Leat can prefer the one with the stronger margin or better stock position. The same approach can select a free-item reward the customer values without creating unnecessary cost.
Recommendations can appear in an e-commerce checkout, member portal, website widget, email, wallet notification, or another connected customer experience. Unified Execution delivers the selected product through the appropriate surface. Results can then show whether the recommendation was accepted, what it added to the basket, and whether the customer continued buying from the recommended category.
Recommendation catalog scoping
Teams can choose which products and categories may be recommended and exclude restricted, regulated, or unwanted items.
Stock and margin data upkeep
Current POS and e-commerce data helps prevent unavailable or unprofitable products from being recommended.
Channel configuration
Recommendations can be enabled by channel, with control over where they appear and how many products are shown.
Promotion priority setting
A product or category can receive temporary priority for a launch or season while remaining relevant to the individual customer.
Reward item review
Teams can review which products are selected as free rewards and confirm that customer value and business cost remain balanced.
Attachment and uplift reports
Add-on rates, additional basket value, measured uplift, and repeat category purchases can be compared across customer groups and channels.
