The blank compose window
Personal outreach at volume means a fresh draft for every contact, every time. Most signals get politely ignored because the writing is the bottleneck.
Wave · AI Email Writer
Generic AI writers know your brand and nothing about this buyer. Wave already knows what each person read, what they should see next, and how they like to hear from you, so the draft arrives grounded in their actual history, waiting for your approval.
Quick answer
The AI Email Writer produces two kinds of drafts. The per-person recap email summarizes what a contact recently read on your site and nudges them toward their best next item, subject line included. Per-audience content framing pitches one asset differently by the reader's role and buying stage. Every draft is grounded in real consumption history, is off by default, is enabled per tenant, and is a starting point a human reviews and approves. Wave never sends.
Key capabilities
Last updated: July 2026
Why teams need this
Intent tools, scores, and recommendations all end the same way: a human has to sit down and write the message. That step is real work, per contact, every time, and it is why insights never become sends.
Personal outreach at volume means a fresh draft for every contact, every time. Most signals get politely ignored because the writing is the bottleneck.
A first-name merge field dropped into one template reads like a mail merge because it is one. Buyers notice the difference between their name and their history.
A prompt-driven writer produces plausible copy about your product. It cannot recap what this person read last week, because it has no idea.
How Wave does it
Wave's writer is the last link in a chain: committee mapping, consumption tracking, next-best-content, and channel and timing predictions all feed the draft.
Before a word is written, Wave has mapped the buying committee, tracked what each person consumed, picked their best next asset, and read their channel and timing preferences.
The recap draft draws on that person's recently read items and their recommended next item. The framing draft draws on one asset plus the audience's role and stage. Clean facts reach the model: titles, summaries, role, stage. Raw page contents never do.
Accept, edit, or discard, with version history kept. Nothing is sent and nothing is pushed into your CRM or marketing platform. You keep the send button.
When a person's recommended next item changes, the recap re-drafts, so the copy your marketer opens is always built on current reading history.
Where it fits
The words are only as good as the facts behind them. Wave's writer inherits every fact the platform has already established about each buyer.
Committee mapping, content intelligence, next-best-content, and channel and cadence predictions all feed the writer, which is why its drafts read like a rep who did the homework and a bolt-on AI writer reads like a mail merge. Every call is spend-capped per tenant, logged with personal details scrubbed, and confined to your own tenant's data. The capability is off by default, enabled per tenant from the Wave console, and its drafts stay inside Wave until a person decides otherwise.
Works alongside Play Engine, Next Best Action Engine, Buying Group Intelligence.
Why Wave is different
Anyone can bolt on a text generator. The difference is what the writer knows before it writes: the committee, the reading history, the next-best pick, and the channel and timing fit.
FAQ
A Wave capability that drafts email copy grounded in each person's real consumption history. It produces per-person recap drafts that summarize what a contact recently read and point to their best next item, and per-audience framing that pitches one asset differently by role and buying stage. A human approves every draft, and Wave never sends.
The per-person recap email: a subject line plus a short body recapping what that person recently read, with a nudge to their recommended next item. And per-audience content framing: a subject line plus a short blurb that pitches one content piece differently depending on the reader's role and stage.
First-party behavior. The recap draws on that person's most recently read items and their recommended next item. The framing draws on the content piece plus the audience's role and buying stage. No generic templates and no guessing.
No. Drafts are internal writing aids. Nothing is sent and nothing is pushed into your CRM or marketing platform. Your team reviews, edits, and approves every word, and you keep the send button.
The recap draft refreshes itself so it never goes stale. What your marketer opens is always built against the person's current reading history and current best next item.
A few clean facts: titles, summaries, the person's role, and their stage. Raw page contents never reach the model. Every call is spend-capped per tenant, logged with personal details scrubbed, and confined to your own tenant's data.
No. It is off by default and enabled per tenant from the Wave console, so the capability arrives on your terms.
Book a 20-minute walkthrough. We will open a real contact record and show the recap draft Wave writes from that person's actual reading history.
See it on your data
Book a 20-minute walkthrough. We will open a real record and show the recap Wave drafts from that person's actual reading history.
Request a demo