منصة ولاء وإدارة علاقات العملاء بالذكاء الاصطناعي للمطاعم
يتنبأ بالعملاء المعرضين للتوقف عن الزيارة ويرسل عروضاً مخصصة تلقائياً عبر واتساب قبل أن يتوقفوا
Overview
Built for a US-based casual dining chain with 34 locations. A loyalty programme nobody used — generic email offers with 6% redemption — was rebuilt as an AI-driven CRM.
The Challenge
The existing loyalty programme sent the same generic email offer to every customer, regardless of visit frequency or preference, and redemption sat at 6%.
What We Built
A CRM that tracks visit frequency per customer, predicts churn three weeks in advance, and sends personalised offers through the customer's preferred channel — timed to the moment they're most likely to respond.
Customizable by Design
Offer rules, loyalty tiers, churn thresholds, and messaging channels are all configurable per brand and per location, with the menu and promotions catalog plugged in rather than hard-coded. The AI works both ways: owners get predictable visit and churn forecasting per location, while customers get offers personalised to their own visit pattern instead of a generic blast.
What Sets This Apart
- •A true closed loop — predict, send, measure, retrain — running hands-off, instead of a churn dashboard someone has to act on manually
- •WhatsApp-native engagement, built for how customers in the region actually communicate, not an email-first tool with WhatsApp bolted on
- •Native ZATCA e-invoicing (FATOORAH) compliance, so franchises don't need a second system just to stay compliant
- •Location- and shift-level churn diagnostics for multi-unit owners, not just an aggregate chain-wide number
- •Reports incremental revenue against a control group, not just redemption rate — proof the spend actually worked
Results
- •67% increase in repeat visit rate within 6 months
- •44% offer redemption rate vs. 6% industry average
- •3-week advance churn prediction at 84% accuracy
- •80% of customer communications fully automated
Tech Stack
النتائج
زيادة الزيارات المتكررة 67% · معدل استخدام العروض 44% · دقة التنبؤ بالتسرب 84%