What is next best conversation personalization?

Next best conversation personalization uses a customer’s current context, behavior, journey stage, and explicit preferences to select the most useful interaction at that moment. Unlike campaign targeting, it begins with triggers such as onboarding, repeated overages, service dissatisfaction, trial expiration, or attrition risk, then coordinates the appropriate message, offer, education, or service response.

How should a company start an AI personalization program?

A company should begin with one valuable customer problem and a measurable proof of concept, not a company-wide technology overhaul. A personalized onboarding or educational experience can test available data, customer response, operational coordination, and economics. Aetna’s SundaySky pilot demonstrated this approach by reducing call-center volume by more than 20% while improving satisfaction and engagement.

What is zero-party data, and how can brands collect it?

Zero-party data is information customers intentionally provide, including preferences, goals, trade-offs, and feedback. Brands can collect it through onboarding questions, loyalty profiles, quizzes, ratings, and gamified assessments. Marriott Bonvoy asks about vacation plans and welcome gifts, while SonderMind uses exercises across 12 dimensions to improve therapist matching and track progress.

How does AI help integrate customer data?

AI can inspect schemas in separate databases, generate transformation code, and standardize information from billing, sales, service, marketing, and regional systems. Narrative’s Rosetta supports this process for companies including Pepsi across 40 geographies, The Trade Desk, and Snowflake. AI can reduce manual integration work, but it cannot repair customer information that teams never captured or governed consistently.

How much should marketers invest in testing personalization?

Dave recommends allocating roughly 10% to 20% of the marketing budget to new tests that improve personalization over time. Teams should move beyond average A/B winners when they have sufficient scale, using multivariate testing to determine which content, offer, timing, and journey work for different customer types. Shopify tested approximately 500 message combinations with OfferFit.

How can brands personalize without undermining privacy?

Brands should use customer data for an evident benefit, obtain appropriate consent, secure the information, and test sensitive experiences before scaling them. BCG research cited by Dave found that 70% of U.S. consumers gave more business to companies that used their data helpfully, while two-thirds stopped buying from companies they believed used it inappropriately.

Build Personalization Around the Next Best Conversation

For marketing leaders building AI-enabled personalization, this session explains how to turn customer data into relevant, valuable interactions at scale.