The 2026 State of Customer Engagement Report: Are AI Filters Already Deciding Who Sees Your Messages?
Six months ago, most marketing teams had AI filtering on their "things to watch" list. Then iOS 26 shipped expanded notification summaries, Gmail started routing messages through Gemini before users ever open their inbox, and the "things to watch" list suddenly became much more immediate.
A new layer of infrastructure, AI filtering, is making decisions about your messages before your customers ever do. How much of your audience is actually seeing what you send? Which channel still has the biggest impact on long-term retention? Is a 9x lift in engagement actually possible, and what's the one variable that explains it?
The 2026 State of Customer Engagement Report is officially out, and aims to equip senders with answers to all of those questions. It's built on aggregated data from the OneSignal ecosystem, billions of messages sent across push, in-app, email, SMS, and RCS, combined with a survey of marketers and product teams across industries.
Download the Report
What follows are the stats that shaped this year's findings, and what they mean for the way you build in 2026.
48% of marketers are already concerned about AI- filtering deciding which messages get seen. 17% say it's already impacting deliverability.
→ What it means
The old assumption that a sent message is a seen message no longer holds. Teams that are still measuring success purely by send volume or open rate are optimizing for a metric that's increasingly disconnected from actual engagement.
The shift this report calls out isn't "should you care about AI filtering?" It's how quickly you build messaging strategies that hold up under it.
Behavior-triggered messages outperform standard sends by 4–9x on CTR.
→ What it means
This is the widest performance gap in this year's data, and it isn't close. A message that arrives because something just happened (a search, a purchase, an abandoned step) earns attention in a way that a scheduled broadcast doesn't, because it maps to a moment the user is already in.
In an AI-filtered world, this structural advantage only grows: messages tied to real behavior signals are more likely to clear OS-level prioritization filters, because they look like what they are: immediately relevant. Teams still running mostly scheduled sends have a clear, high-ROI optimization in front of them.
75.8% of teams say push has the greatest impact on long-term retention in the first 30 days after install, more than email and in-app combined.
→ What it means
The first month is still the fight that matters most. Push isn't losing ground as a retention channel… but the standard for what a good push looks like has risen. Before a notification reaches a user, operating systems are increasingly summarizing, grouping, and suppressing. Which means the first 30 days have become about building an “attention credit score”: demonstrating to the OS, and to the user, that your notifications consistently lead somewhere worth going.
45% of mobile teams are actively exploring or planning RCS adoption in 2026.
→ What it means
RCS is no longer an "eventually" conversation. What the report makes clear is that the teams moving on this now are first in line to earn this new relationship surface. RCS turns the inbox from a place where messages arrive into a place where customers now enjoy a conversational and app-like experience. The expectations for text marketing are about to soar.
63% of teams using automated Journeys report better results with fewer messages sent. 78% report higher revenue from users included in Journey flows.
→ What it means
Automated sequencing is how teams at any size get compounding returns without compounding effort. The data here runs counter to the instinct to "send more." The teams reporting better performance aren't sending more at all; they're sending in the right order, stopping when the job is done, and letting behavior drive when the next message goes out.
The lift is consistent: among teams that can tie Journeys to revenue, the majority are seeing double-digit improvements.
65% of teams say they're experimenting with AI in their messaging workflows. Only 8% have fully operationalized it.
→ What it means
The experimentation phase is widespread, but real integration is rare. That gap makes sense, since AI is arriving faster than most teams can build the governance around it.
The clearest early wins are in low-stakes, high-frequency work: subject line testing, copy variants, send-time optimization. The longer-term promise, AI identifying when a message is warranted versus when silence is better, is where the real leverage sits, and most teams aren't there yet.
Even more insights in the full report…
The 2026 State of Customer Engagement Report not only includes expanded tactics and frameworks for the above findings but also industry-level benchmarks for:
- Push opt-in rates
- Session duration
- 1/7/30-day retention
- Annual churn rates