A Philippine telco transforms Customer Value Management with AI/ML-driven omni-channel campaigns
How OneByZero designed an ML-powered Customer Value Management framework integrated with Adobe's omni-channel platform, delivering real-time, personalised next-best-offer decisioning on AWS at national scale.
Built on AWS SageMaker with Adobe omni-channel integration
8×
improvement in campaign efficiency, from 7% to 55%
6×
increase in daily customer coverage, to 6M per day
+10%
ARPU uplift per customer
Overview
About the customer
A leading telecommunications and digital services provider in the Philippines faced mounting challenges in a highly competitive industry where personalisation and targeted campaigns are critical. Its Customer Value Management (CVM) relied on manual, gut-driven campaign designs with limited experimentation, basic segmentation, and slow reporting, resulting in low campaign efficiency, poor personalisation, and limited revenue impact. Scaling was further constrained by on-premises infrastructure, which limited experimentation, slowed deployment, and capped subscriber reach.
The challenge
Modernising CVM in a competitive, scale-constrained market
Without modernisation, the client risked revenue leakage, higher churn, declining ARPU, and structural disadvantage versus digitally mature competitors. Four priorities defined the transformation.
01
Manual, low-efficiency CVM
Campaigns were gut-driven with limited experimentation, basic segmentation, and slow reporting, capping revenue impact.
02
Fragmented channel engagement
The operator needed to use omni-channel communications, SMS, push, USSD, and social, effectively and in concert.
03
Limited efficiency and adoption
Campaign efficiency, adoption, and revenue growth all needed to improve at national scale.
04
On-prem scalability limits
On-premises infrastructure limited experimentation, slowed deployment, and capped subscriber reach.
The solution
An ML-powered CVM framework on AWS
The client partnered with OneByZero to design and deploy a comprehensive ML-powered CVM framework integrated with the Adobe omni-channel platform, combining omni-channel communication with AI/ML decisioning to deliver real-time, personalised offers.
#01
Adobe omni-channel integration
Reach customers across every channel.
- SMS, push notifications, USSD, and social media enablement
- Flexible response generation (spiels) to experiment with tone and messaging
#02
Ensemble next-best-offer models
Offers optimised for both customer interest and business revenue.
- Combines customer value (propensity, persuadability) and business value (ARPU uplift) scores
- Algorithms: XGBoost, PyTorch TabNet, and persuadability models
#03
A/B testing on AWS SageMaker
Systematic experimentation across customer segments.
- Expanded from 10 to 60 configurations, refined to 5 best-performing models
- Tested across new, long-tenure, and high-value segments
#04
End-to-end ML pipeline
Production MLOps from training to inference.
- Preprocessing, training, model registration/versioning, deployment, and inference
- Real-time inference optimised for <200ms latency SLA
- Auto-scaling endpoints handling up to 600 requests/sec
#05
Hybrid cloud deployment
AWS-native, integrated with on-prem systems.
- AWS: SageMaker, Lambda, API Gateway, DynamoDB, S3, Glue
- Seamless integration with on-prem data and marketing automation
#06
Operational enablement
Built for oversight and scale.
- Monitoring dashboards and load testing for high concurrency
- L1 support and training for client IT teams
AWS services used
A hybrid, AWS-native ML stack
The solution runs on AWS-native services across data, training, inference, and serving, integrated with the client's on-premises systems for a hybrid deployment.
Amazon SageMaker
Feature Store, model training, registry, and scalable real-time inference endpoints for the next-best-offer models.
AWS Lambda
Serverless linear model that invokes inference endpoints and computes the weighted Customer Value and Business Value score.
Amazon DynamoDB
Low-latency store for Business Value metrics used in real-time decisioning.
Amazon API Gateway
Managed entry point for external inference requests, routing them to Lambda.
Amazon S3
Foundational data lake for ingested data and ML artifacts.
AWS Glue
Crawler jobs that extract and orchestrate data into the S3 data lake.
Delivered by OneByZero, an AWS partner, integrating AWS-native ML services with the client's existing on-premises data and marketing automation platforms.
Outcome
Personalised engagement at national scale
8×
campaign efficiency, from 7% to 55%
6×
daily customer coverage, to 6M/day
+10%
ARPU uplift per customer
The initiative delivered measurable improvements across reach, efficiency, and revenue, transforming CVM from a manual, gut-driven function into a data-driven, AI-enabled capability.
- Customer coverage grew 6×, from 1M to 6M per day
- Net take-up rate doubled with ML-driven decisioning
- Campaign efficiency rose 8×, from 7% to 55%
- ARPU up 10% per customer; ticket size up 30%
- Inference latency cut 72%, from >500ms to <200ms
- Subscriber base scaled 3×, from 10M to 30M+ on cloud
About OneByZero
OneByZero is a frontier Systems Integrator specialising in AI Coworker design, build, and deployment for regulated enterprises across financial services, telecommunications, and retail. We operate across ASEAN, India, ANZ, and Japan, with a presence in the United States, combining domain expertise in regulated industries with deep AI engineering capability.
HQ Singapore · Tech HQ USA · Australia · India · Indonesia · Malaysia · Philippines · Thailand · Vietnam
