E-commerce has spent years refining its ability to predict customer behavior, but the next competitive frontier is deciding what to do with those predictions while the customer is still engaged. That shift is beginning to redefine the role of e-commerce technology, moving from historical analysis and fixed segments toward systems that interpret real-time signals and determine the next best action.
Adobe's 2025 retail research found that 45% of retailers were using generative AI for next-best-action recommendations and 46% were using AI to personalize experiences based on real-time customer behavior. Yet 41% said fragmented data was preventing them from delivering real-time personalization. Adobe's 2026 research underscores the tension: 56% of organizations identified more personalized customer experiences as a top AI investment goal, and 80% said highly personalized experiences that anticipate needs in real time will define breakthrough customer experience over the next two to three years. However, only 44% said their data quality and accessibility were adequate for AI, and 75% identified data integration and quality as a major challenge.
"The industry has spent years getting better at predicting customer behavior," said Ersan Erturk, founder of Peloran. "The harder question is what to do with that prediction when the customer, the product and the commercial context are changing at the same time."
A predictive model can estimate that a shopper is likely to purchase, hesitate or leave, but that prediction alone does not determine the best response. A customer may show strong purchase intent while also being price-sensitive. Product availability or sales performance may have shifted. A returning visitor who repeatedly views a product might warrant different treatment than a first-time visitor arriving through a promotion. The decision depends on more than predicted intent; it requires understanding what the customer is doing, what the product is doing, what is happening commercially, and what has already occurred during the journey.
Peloran is addressing this problem with its commerce intelligence platform, which brings customer behavior, product activity, sales performance and pricing context into a common intelligence layer. The platform generates behavioral intelligence including purchase intent, hesitation, value orientation, discount sensitivity and customer value, then connects those signals to customer journeys and automated actions. For merchants evaluating a marketing automation tool, this creates an approach where automation responds to a broader set of signals rather than relying solely on predefined triggers.
The move toward real-time decisioning creates a fundamental data problem. Retailers have accumulated customer, product, transaction, marketing and operational data across systems often designed to work independently. Adobe's 2025 research found that 41% of retailers said fragmented data was preventing real-time personalization, while 35% said it was creating inconsistent experiences across channels. Adobe's 2026 research found that only 39% of organizations have a shared customer data platform capable of supporting widespread agentic AI adoption. The challenge is no longer simply collecting more data; it is connecting the right signals quickly enough to make a useful decision.
Peloran connects its intelligence layer to customer journeys and marketing automation, allowing merchants to act across storefront experiences, email, WhatsApp, web push, advertising and product experiences. A high-intent visitor showing hesitation can enter a different journey from a price-sensitive shopper. Product and pricing context can influence whether an intervention makes sense, and a customer who does not need an intervention can be left alone. For merchants considering shopify marketing automation, the approach connects automated marketing decisions with behavioral and commercial signals.
Real-time action also creates a measurement challenge. A customer may purchase after receiving an offer, but that does not necessarily mean the offer caused the purchase. Peloran uses exposed and holdout groups to compare outcomes and estimate incremental revenue from interventions, changing the question from "Did the customer buy?" to "Did the intervention create additional revenue?" As more commerce decisions become automated, measuring incremental impact becomes as important as deciding which action to take.
The broader shift is not from prediction to no prediction. It is from prediction alone to prediction-informed decisioning—systems designed to continuously determine what should happen next. For merchants using customer journey mapping tools, this approach allows journeys to incorporate changing customer and commercial context rather than functioning only as fixed sequences.
