Mobile Dev Memo Podcast Season 7, Episode 10: Deploying AI personalization at scale
Mar 17, 2026
Christina Augustine, COO at Bloomreach and former software engineer, leads AI-enabled personalization and agentic commerce. She traces the move from predictive models to generative agents. Short takes cover real-time one-to-one personalization vs cohorts, channel trade-offs like SMS versus email, data quality as the main bottleneck, conversational shopping benefits, and the rise of agent-to-agent personalization.
AI Snips
Chapters
Transcript
Episode notes
Multi-Channel Signals Replace Basic Cohorts
- AI personalization now combines signals across email, SMS, social, search and onsite to treat customers as individuals rather than cohort labels.
- Christina Augustine explains using multi-channel behavior (e.g., time on site thresholds) to trigger different experiences like pop-up agents for shoppers who struggle.
Send SMS Only To High Value Segments
- Prioritize personalization by channel based on cost and impact, sending expensive messages like SMS only to high-value, responsive segments.
- Christina Augustine recommends matching channel strategy to business goals (top-line vs bottom-line) to avoid overspending.
Data Quality Is The Main Bottleneck
- Data quality is the primary bottleneck for scalable AI personalization because AI is only as good as the input signals.
- Christina Augustine notes platforms retaining native data makes it harder to know search intent and optimize channels like AEO.