Yum! Brands
generation of personalized CRM communications at scale (genAI on first-party data)
In 2025, Yum! Brands (Taco Bell, KFC, Pizza Hut) sent more than 200 million AI-generated marketing communications, up to five times more incremental than traditional approaches, relying on its Red360 permissioned data program (140 million names) and its in-house Byte by Yum! platform.
Key points
- Generation of personalized CRM communications at scale (offers, email subject lines, push messages).
- In-house Byte by Yum! platform, Red360 first-party data, and OfferFit (Braze).
- More than 200 million AI-generated communications, up to 5x more incremental.
- Evidence level A, confirmed active status.
Objective
Turn a base of more than 140 million permissioned names into incremental sales by generating personalized marketing communications at scale (offers, email subject lines, push messages) rather than broadcasting the same campaigns to everyone.
The deployment
Yum! Brands relies on its Red360 first-party data program (more than 140 million qualified, permissioned names, with transaction history) and on its in-house Byte by Yum! technology platform to produce individualized marketing communications at large scale. In 2025, the group states it sent more than 200 million AI-generated communications, up to five times more incremental than traditional approaches. Concretely, the AI selects and writes the offers, email subject lines, and push messages per member, where the marketing team previously sent uniform campaigns. Yum laid out eleven AI marketing use cases, of which seven were active in the second quarter of 2025. For email and SMS personalization, the group relies on OfferFit, an offer-optimization component acquired by Braze. This setup is part of a broader digital shift: in Q2 2025, Yum's digital sales mix reached 57% (7 points more year-on-year) and digital sales grew 18%. This case covers the generation of personalized CRM communications, distinct from the deployment of voice AI at the Taco Bell drive-thru, which falls under another pattern.
Results Proof A
Figures announced by the CEO and CFO in the Q2 2025 earnings call (verbatim transcript), corroborated by the established marketing press and by continuity signals in the following quarters (Q4 2025, Q1 2026).
How it works
Documented architectureThe stack in detail
- plateforme Byte by Yum! In-house multi-brand technology platform (digital ordering, operations, personalization, AI), present in around 25,000 restaurants
- plateforme Red360 First-party permissioned consumer data program (more than 140 million names with transaction history)
- outil OfferFit Self-learning offer optimization for email and SMS personalization, acquired by Braze
How it runs, concretely
For ops teams-
1Feed the permissioned data base data team
Collect and unify in Red360 the permissioned names and transaction history, primarily via Taco Bell, which provides about half.
-
2Define the marketing use cases marketing / data team
Frame the AI marketing use cases (offers, email subject lines, push messages, in-app recommendations) and decide which go to production.
-
3Generate and target the communications AI
The AI writes and selects per member the offer and message, with the OfferFit component continuously optimizing which offer to send to maximize incrementality.
-
4Distribute on the CRM channels AI / Byte platform
Send via email, SMS, push, and in-app recommendations on the brands' app.
-
5Measure incrementality and loop back data team
Compare the sales generated to a reference scenario, then feed the system back to adjust offers and frequency.
Incrementality measured per member (sales attributable to a message versus a scenario without any send). Without this measurement and without Red360's permissioned data, the AI no longer knows which offer to push to whom, and targeting falls back to uniform campaigns.
How your customers perceive this type of use
Sourced studiesUn ecart net separe les annonceurs des consommateurs : 77% des annonceurs voient l'IA positivement contre 38% des consommateurs (Yahoo/Publicis, 2024). Les mesures implicites confirment le rejet declare : en EEG, les pubs generees par IA produisent une activation memorielle plus faible que les pubs traditionnelles et sont decrites comme agacantes, ennuyeuses et confuses (NIQ, 2024). La disclosure a un effet ambivalent : elle augmente fortement la confiance quand elle est remarquee (Yahoo/Publicis), mais 27% des jeunes consommateurs disent faire moins confiance a une entreprise dont la pub est creee par IA (IAB, 2024).
Acceptance conditions
- Une disclosure visible : quand la mention IA est remarquee, la confiance globale envers l'entreprise augmente de 96% (Yahoo/Publicis 2024)
- Une qualite visuelle suffisante : les visuels IA de basse qualite augmentent l'effort cognitif et distraient du message (NIQ 2024)
Red lines
- Le contenu IA non declare puis identifie : 72% des consommateurs disent que l'IA rend l'authenticite difficile a etablir (Yahoo/Publicis 2024) et les marques utilisant des pubs IA sont plus souvent jugees inauthentiques ou non ethiques par les consommateurs que par les dirigeants (IAB 2024)
- Les mannequins et personnes generes par IA : 46% des consommateurs n'en veulent pas dans la publicite, l'inquietude premiere etant les standards de beaute irrealistes (Attest 2025)
Sources: Yahoo / Publicis Media (terrain Ebco) 2024 · IAB (avec Attest) 2024 · NIQ (NielsenIQ) 2024 · Attest 2025
How to replicate
Inference, not sourcedData prerequisites
- A permissioned, unified contact base linked to transaction history per person
- A stable customer identifier to attach sends to actual purchases and measure incrementality
- A control-group measurement setup to distinguish incremental sales from sales that would have occurred without a message
Org prerequisites
- A data or CRM team that owns incrementality measurement, not just sends
- A governance framework for AI marketing use cases with a clear sorting between pilot and production
- Legal basis and GDPR consent for personalization, and transparency about AI-generated content
Possible stack
- A CDP or first-party customer warehouse
- A content-generation component (offers, subject lines, messages) by LLM
- A self-learning offer-optimization engine (OfferFit type) on the email, SMS, and push channels
- A CRM platform and a mobile app for distribution and in-app recommendations
First step: Set up incrementality measurement with a control group on an existing channel (email or push), before even automating generation: it is the signal that drives the whole setup.
The same pattern, proven in other industries
Allstate
genAI drafting of customer communications with human review
Google for Education
genAI assistant integrated into an institutional education suite, admin-managed
Spotify
genAI voice DJ (synthetic voice plus generated commentary)
Sources
- S1 Yum! Brands (YUM) Q2 2025 Earnings Call Transcript Primary archive pending
- S2 How Yum's AI factory supercharges marketing at Taco Bell and beyond Established press archive pending
- S3 Yum Brands Battles Rival Restaurants With Proprietary AI Secondary archive pending
An error, newer info, a source?
This page lives on its accuracy. If a figure has moved, if the deployment has changed, or if you have a higher-quality source, tell us. Every sourced correction is verified before publication.