Coca-Cola is using AI to recommend what Malaysian retailers should order and in what quantities. The Perfect Basket feature sits inside Coke Buddy, the company’s self-ordering platform, and analyzes past orders, ordering frequency, seasonality, weather and purchasing patterns among similar businesses.
Coke Buddy supports about 39,000 retail outlets across Malaysia, according to Coca-Cola. Retailers can place orders through an app, website or WhatsApp, receive personalized suggestions and track delivery. Perfect Basket moves those suggestions closer to the purchasing decision by presenting proposed products and quantities before an order is completed.
Recommendations combine local and historical signals
Coke Buddy already used purchase history to suggest products a retailer might reorder. Perfect Basket adds contextual signals, including seasonal conditions, weather and trends among comparable outlets. The final choice remains with the retailer, who can review and change the proposed basket before submitting it.
Fulfilment is handled separately by Coca-Cola Refreshments Malaysia or its suppliers through existing sales and distribution arrangements. Coca-Cola also says its sales teams continue working with retailers. The system is therefore presented as an aid to routine ordering rather than a replacement for either the store owner’s decision or the company’s account relationships.
That division of responsibility matters because the source does not disclose how the recommendation engine resolves conflicting signals, how frequently it updates or how it handles unusual local demand. A proposed quantity may be informed by similar stores and recent weather, but the retailer still has information about neighborhood events, shelf space and cash constraints that is not described in the model inputs.
A campaign produced adoption, not a full performance audit
Coca-Cola disclosed usage data after a Perfect Basket, Perfect Ride campaign that ran from January through April 2026. More than 4,000 Malaysian retailers submitted over 4,500 entries, and the company says 83% of participating outlets adopted Perfect Basket recommendations.
That percentage applies only to retailers who joined the campaign, not all 39,000 outlets using Coke Buddy. Coca-Cola also reports that participants following the recommendations recorded stronger sales-revenue growth than comparable outlets, but it did not disclose the size of the difference or provide detailed results linking individual suggestions to sales and inventory changes.
Several important operating measures are absent from the Malaysian data. The source gives no figures for forecast accuracy, product availability, inventory levels or logistics costs. The campaign demonstrates that many participating retailers were willing to accept suggested baskets; it does not yet establish the feature’s average financial effect across the wider network.
Similar systems show the wider sales strategy
Coca-Cola has deployed suggested-order tools elsewhere in its bottling network. In first-quarter 2024 results, the company said it and its bottling partners had connected nearly eight million customers to business-to-business platforms worldwide, while AI-assisted ordering reached more than three million outlets in Latin America.
Earlier pilots provide additional, but separate, performance claims. Coca-Cola said retailers receiving AI product recommendations based on previous orders and market data were more than 30% likelier to buy the suggested stock-keeping units. Those results did not come from Perfect Basket in Malaysia. In another three-country demand-prediction pilot using sales, weather and location data, the company reported sales 7% to 8% higher than at outlets not using the algorithm.
Coca-Cola has linked these systems to a change in sales work. AI-generated orders can reduce the time pre-sales employees spend collecting routine requests, allowing more attention to account development, while stores retain approval of what they buy.
Perfect Basket remains available after the campaign, and Coca-Cola says it will continue refining Coke Buddy with retailer feedback and data. The next evidence will need to go beyond campaign participation: performance across the full outlet base, measurable inventory outcomes and the frequency with which retailers override the suggestions would show how effectively automated recommendations fit real purchasing decisions.
