Wave · Engagement outcomes
Reach every buyer the way they actually want to be reached. Send-time tools learn from your outbound history, so they are blind until you have mailed someone for months. Wave learns from how each person arrives and what they consume: UTMs, referrers, AI-assistant traffic, and content behavior. Channel, format, and timing exist from the first inbound touch.
Quick answer
Predict Every Buyer is Wave's prediction capability: the channel each person responds to, the content format they consume, and the timing that fits their behavior, predicted per person and written to your CRM with a confidence grade. It trains on inbound signals, UTMs, referrers, AI-assistant arrivals, and content consumption, so predictions exist even for contacts with zero send history.
What this capability includes
Last updated: July 2026
Why it compounds
This is the preferences beat of the Wave journey: from the content and people Wave already maps, it learns how each person likes to engage. Each feature inside this capability answers one question about a person. Stacked on the same person, they answer the one that matters: how do I reach this specific human?
The compounding is structural. Channel Affinity and AI Referral Intelligence read the same arrival stream, so the channel prediction already understands AI-era traffic. Engagement Format reads consumption on the same person, so the format matches the channel it travels through. Cadence Affinity times the whole thing. A point tool can produce one of these numbers inside its own silo. It cannot produce a playbook, because it never sees the other three answers for the same human being.
Inside this capability
Each one stands alone. Stacked on the same person, they stop being metrics and become instructions.
Predicts the channel
Learns the channel each person actually responds to from how they arrive: UTMs, referrers, and inbound touches. Wave can predict a channel you have never used on that person, because sends are not the training data.
Deep dive →Sees AI-era arrivals
Classifies traffic arriving from ChatGPT, Perplexity, Claude, Gemini, and Copilot as a first-class channel instead of letting it vanish into direct. The buyers your analytics cannot see become buyers you can predict.
Deep dive →Predicts the format
Learns whether each person is a webinar person, a whitepaper person, or a video person from what they consume, not what your MAP happened to send them.
Deep dive →Predicts the timing
Predicts the send window and touch rhythm that fits each person's behavior, including the people your ESP has never mailed and therefore cannot optimize.
Deep dive →Why one platform
Send-time optimizers and channel reports train on your outbound history. The answer is circular: they can only recommend what you already do.
FAQ
Four Wave features working on the same person: Channel Affinity predicts the channel, Engagement Format predicts the content format, Cadence Affinity predicts the timing, and AI Referral Intelligence captures arrivals from assistants like ChatGPT and Perplexity so the channel model understands AI-era traffic. Every prediction carries a confidence grade and writes to your CRM.
Because sends are not the training data. Wave learns from inbound behavior: the UTMs and referrers a person arrives with, the AI-assistant traffic they ride in on, and the content they consume. That signal exists whether or not you have ever mailed them, which is why Wave can recommend a channel your team has never tried on that person.
First-party inbound signals from your own systems: UTM parameters, referrer data, AI-assistant referrals, and engagement history from your connected stack, including imported CSV and warehouse history. One model per tenant. Your data never trains another company's predictions.
Wave declines to guess. When a person's history is too thin to support a real prediction, Wave predicts nothing rather than emitting low-confidence noise. As inbound signal accumulates, predictions appear with their confidence grades.
On each contact inside Wave, side by side across channel, format, and timing, and as native HubSpot contact properties your workflows and segments can route on. Writeback is off by default, opt-in per tenant, and every write is logged with previous value and new value.
Your ESP optimizes sends using its own send history, so it is silent on anyone you have not mailed and blind to every channel it does not operate. Wave predicts channel, format, and timing together from inbound behavior, which covers your whole database, including the contacts you have never emailed.
See it on your data
Book a 20-minute walkthrough. We will run Wave against your inbound history and show the channel, format, and timing predictions it would write to your real contacts.
Request a demo