AI-powered product recommendation engine for self-ordering kiosks

- KFC Vietnam operates 250+ restaurants with self-ordering kiosks serving millions of customers monthly. - Kiosks currently display static menus with no personalization. Customers are not prompted with relevant upsell or cross-sell suggestions based on order context such as time of day, current selection, or combo affinity. - This creates missed revenue opportunities and a suboptimal ordering experience. KFC estimates 15-20% potential uplift in average order value if relevant recommendations are surfaced at the right moment in the ordering journey.

- Increase average order value by 10-15%. - Improve relevance during the kiosk ordering journey. - Give marketing and operations a data-driven upsell channel.

Kiosks currently display a static You may also like section manually curated by the marketing team and updated monthly. There is no real-time personalization; recommendations are identical for all customers regardless of time of day, current selections, or location.

- Kiosks currently display a static You may also like section manually curated by the marketing team and updated monthly. - There is no real-time personalization; recommendations are identical for all customers regardless of time of day, current selections, or location.

- End customers across 250+ KFC Vietnam restaurants get a faster and more relevant ordering experience. - Marketing gains a dynamic, data-driven upsell channel replacing manual curation. - Operations and restaurant teams can increase AOV without additional staff effort.

POS transaction records: item-level order history across 250+ restaurants. Menu catalog: structured product data including categories, pricing, combos, and nutritional info in a relational database. Contextual signals: restaurant location, time of day, day of week, and current promotional calendar.

The recommendation engine must integrate with the existing self-ordering kiosk platform. --- *Track: F&B | Sponsor: KFC | Slug: fnb:kiosk-recommendations | Source: Agentic AI Build Week 2026 — Production snapshot 2026-08-11*