Uni-President BU Integration PoC
📅Published 2026-05-14·🔄Updated 2026-07-02·⏱4 min read
"BU integration innovation PoC that ties OPENPOINT members, food manufacturing, 7-Eleven, Carrefour, Starbucks, and logistics data together with external trend, weather, economic, and competitor signals — solved through ontology + Agentic AI."
For Uni-President (統一企業), this PoC unifies its own BUs (Uni-President Food manufacturing + 7-Eleven Taiwan + Carrefour + Starbucks + Mister Donut + KFC) and the cross-BU, cross-channel behavior of 14M OPENPOINT members into a single Knowledge Graph. Uses publicly available information only.
🚀 Interactive Demo
👉 Open Demo Page (new window)
Experience the core UIs hands-on: 11 scenarios + cross-BU OPENPOINT member journey, cold-chain SLA, and own-SKU sell-through.
1. PoC One-Liner
| Axis | Content |
|---|
| For Whom | 5 departments — Integrated Marketing, CMI, D&A, OPENPOINT, Manufacturing & Logistics |
| What | OPENPOINT + own SKUs + 7ELE/Carrefour/Starbucks transactions + cold chain + 4 external sources |
| Differentiator | Cross-BU member journey + own-SKU sell-through + cold-chain SLA |
| BU | Content |
|---|
| Uni-President Food (統一食品, manufacturing) | Own-manufactured SKUs such as Uni-Noodle (統一麵), Mai-Hsiang (麥香) beverages, dairy |
| 7-Eleven Taiwan | 6,800 stores (Taiwan's #1 convenience store) |
| Carrefour Taiwan (acquired in '23) | Hypermarkets + minimarts, ~330 stores |
| Starbucks Taiwan | Joint-venture operation |
| Mister Donut / KFC / 21Century | Foodservice BUs |
| President Transnet | Own logistics (cold chain) |
| OPENPOINT | 14M members (BU-integrated membership) |
3. 5-Department Personas
| Code | Department | KPI |
|---|
| P1 | Integrated Marketing (cross-BU) | OPENPOINT campaign ROAS · cross-BU usage |
| P2 | Consumer & Trend Insights | LTV · personas · 7ELE↔Carrefour crossover |
| P3 | D&A Platform | Model accuracy · cross-BU ETL SLA |
| P4 | Membership · OPENPOINT | Tier transitions · earning/redemption rates · NPS |
| P5 | Manufacturing · Logistics · Stores | Uni-President SKU sell-through · store inventory · cold chain |
4. 11-Scenario Summary
| # | Scenario | Data Mix |
|---|
| S1 | Natural-language semantic search | Own SKUs + members + external reviews |
| S2 | 5-department persona chatbot | All tools invoked autonomously |
| S3 | Category · BU insight cards | Per-BU GMV + external |
| S4 | Persona matching + clustering | RFM + categories + social |
| S5 | Omnichannel campaign ROAS simulation | Campaigns · OPENPOINT |
| S6 | External signal fusion | 4 sources |
| S7 | Omnichannel member journey | 7ELE→Carrefour→Starbucks |
| S8 | Guardrails | PDPA compliance + minors |
| S9-U ⭐ | Cross-BU OPENPOINT member journey | One member's cross-BU behavior |
| S10-U ⭐ | Own-manufactured SKU vs own-channel sell-through | Uni-President SKU turnover at 7ELE/Carrefour |
| S11-U ⭐ | Cold-chain · logistics SLA + store-order optimization | Cold chain vs store inventory + outdoor temperature |
5. AWS Stack
Same as LG H&H + reinforced cross-BU ETL via Glue, TimeStream for cold chain, ~900K edges (BU diversity).
6. Top-5 Demo (30 min)
| Order | Scenario | Duration |
|---|
| 1 | S1 Semantic search | 4 min |
| 2 | S9-U ⭐ Cross-BU OPENPOINT journey | 6 min (PoC decisive moment) |
| 3 | S10-U ⭐ Own-SKU sell-through | 4 min |
| 4 | S11-U ⭐ Cold-chain SLA | 4 min |
| 5 | S2 Persona chatbot | 4 min |
7. Deliverables