Getting into the inbox used to be enough. Then it had to be at the top. Now it has to survive being read by AI before your subscriber even sees it. If you missed the first shift, the full picture is in our earlier post on email deliverability after Gmail’s AI changes.
Before your subscriber reads your email, AI reads it first. It generates an email summary, typically one or two sentences, and that’s what many people see before they decide whether to click to read the email at all.
This is not a Gmail-specific story. AI email summaries are now standard across Gmail (Gemini), Microsoft Outlook (Copilot) and Apple Mail (Apple Intelligence). The feature is on by default for most users. Your emails are being summarised whether you have planned for it or not.
This is what it means for your email system and your click through rate.
What AI email summaries actually do
When a subscriber opens an email, AI generates an email summary card that appears above the message content. It pulls from the first readable text in the email, looking specifically for clear value propositions, deadlines, calls to action and factual information.
The AI cannot read images. If your email opens with a hero image, a logo block or a banner with no actual text beneath it, the AI has nothing to work with. The email summary it generates will read something like: “This email contains images.” Source. That will be the first impression to your emails now.
Clear structure, front-loaded copy and a single clear purpose give the AI something useful to work with.
The implications of these AI email summaries for your click through rate are significant. Open rates have increased since AI email summaries launched, but click through rate has dropped. The most likely explanation is that AI is auto-opening emails to generate summaries, which inflates open counts. At the same time, subscribers who get a clear enough email summary feel they have already absorbed the content and do not click through to read more. The reader never gets to your email, so they are never presented with the link to click.
Open rates were already an unreliable metric after Apple Mail Privacy Protection pre-loaded tracking pixels from 2021. AI email summaries make them even less reliable. Click through rate is now the number to watch.
The manage subscriptions problem
Alongside AI email summaries, Gmail has introduced a manage subscriptions view. It lists every sender a subscriber receives email from, ranked by frequency, with a one click unsubscribe option next to each one.
This is a separate feature but it makes the same problem worse. Send too frequently to the wrong people and you will not just generate poor AI email summaries. You will appear at the top of a list that lets subscribers remove you in a single tap.
The brands that appear at the top of that manage subscriptions list are not always sending bad emails. They are sending too many emails to too many people. Volume without relevance is now visible to your subscribers in a way it never was before.
One click unsubscribe has been a sender requirement from Google and Yahoo since early 2024. The manage subscriptions view takes it further by making the decision feel proactive rather than reactive. Subscribers are not just opting out when they are annoyed. They are being invited to audit who they hear from more prominently. One click unsubscribe means the barrier to leaving is effectively zero.
Why this is a systems problem, not a copywriting problem
The immediate response to AI email summaries is to fix the writing.
- Put your most important information at the top of your email.
- Keep emails short.
- Use clear calls to action.
Those things help but they only fix one part of the problem.
Each email needs to be clear enough that AI can generate a useful summary of it. A vague email produces a vague summary. The writing fixes the summary. The system determines whether anyone sees your emails at all.
How often you send emails, who you send to and whether those emails are relevant, that is what shapes your sender reputation and your exposure in the manage subscriptions view.
Good writing improves the email summary. A well-run system earns the right to be seen in the first place.
What to fix. And in what order.
These are the structural changes that make the biggest difference under AI filtering.
Front-load the value.
The first 100 to 200 characters of readable text are what AI uses to generate the email summary. That is roughly two short sentences. If your email opens with “Hi [first name], hope you are having a great week!”, that is what appears. Lead with what the email is actually about.
One email, one purpose.
AI email summaries penalise complexity. An email that covers three topics generates a vague summary that communicates nothing. Focus on one email, one clear action, one call to action. Better structure improves your click through rate.
Segment before you send.
Relevance is now a deliverability signal. An email that is irrelevant to its recipient generates no engagement and no engagement weakens sender reputation. Weakened sender reputation will reduce visibility for future emails. The cycle is continuous. Tighter segmentation is the solution to break this cycle.
Manage your frequency deliberately.
Gmail’s manage subscriptions view shows subscribers exactly how often you email them. If you are sending frequently to an unsegmented list, you are easy to remove with the one click unsubscribe. Review how often you send to each segment, not just how often you send overall.
Know which emails are working and why.
If you do not have a clear picture of which emails generate genuine engagement, which subject lines improve click through rate and which segments are drifting towards inactivity, you are optimising in the dark. Understanding your baseline is not optional anymore, it’s the starting point for everything else.
Start with a clear picture of where you are
AI email summaries have changed what good looks like inside the inbox. But the businesses that will adapt most easily are not the ones that rewrite their subject lines. They are the ones that put a structure in place, and prioritise a well-segmented email list with clear ownership and a consistent approach.
If you are not sure where to start, the Email Health Check is a useful first step. Six questions, five minutes and a clearer picture of where the gaps might be.
AI is not going to stop reading your emails. The question is whether yours are giving it something useful to say.
