There's a new reader sitting between you and every person on your email list. It got there before your customer did, and it's already deciding whether your email is worth showing.
Apple Intelligence is generating AI summaries that replace your preheader text in the inbox. Gmail just launched "Manage subscriptions," which ranks brands by how many emails they send and puts an unsubscribe button right next to each one. And on the filtering side, AI-powered spam detection is evaluating your content quality, engagement history, and even the tone of your writing to decide where your email lands.
This isn't a future problem. It's happening right now, on devices your customers are using today. And if you're still writing and sending emails the same way you were a year ago, the gap between what you're sending and what your customers actually see is growing fast.
What's Actually Happening in the Inbox
Let's break down the three big shifts, because they're all different and they all matter.
Apple Intelligence Email Summaries
On any iPhone 15 or 16 (and M-series Macs) running iOS 18+, Apple Intelligence automatically generates a short AI summary of every email. That summary shows up directly in the inbox, replacing your preheader text before the person even opens the message.
Read that again. Your carefully crafted preheader? Gone. Replaced by whatever Apple's on-device AI thinks your email is about.
The summary is generated from the actual content of your email, not your subject line or preheader. So if your email is a wall of images with minimal text, the AI has almost nothing to work with. If your email buries the main point under three paragraphs of filler, the summary might miss it entirely.
And there's no way to opt out. You can't override it or control what the summary says. The user can toggle it off, but most won't. This is the new default first impression for a huge chunk of your list.
Gmail's "Manage Subscriptions"
In July 2025, Google rolled out "Manage subscriptions" across web, Android, and iOS. It's a centralized hub where Gmail users can see every brand that's emailing them, sorted by who sends the most.
The brand sending the most emails is at the top of the list. Right next to each brand is a one-click unsubscribe button.
Think about that for a second. If you're a DTC brand sending 5-7 emails a week, you're probably near the top of that list for your subscribers. And the unsubscribe button is literally one tap away. No hunting for fine print, no "manage preferences" pages. One click and you're done.
This isn't about punishing email marketers. Adobe's take on it is right: a subscriber who unsubscribes is infinitely better for your sender reputation than one who marks you as spam or silently disengages. But it does mean that high-frequency, low-value sending has a much shorter shelf life than it used to.
AI-Powered Filtering and Engagement Scoring
This is the one most brands aren't paying attention to, and it's probably the most consequential.
Mailbox providers (Gmail, Yahoo, Microsoft, Apple) are using AI to evaluate your emails on a level that goes way beyond spam keywords and authentication checks. They're looking at:
- Engagement patterns over time - not just opens, but clicks, replies, forwards, how long someone spends reading, and how quickly they delete
- Content quality and tone - AI filters now evaluate whether your email reads like something a person would actually want to receive versus mass-produced marketing
- Behavioral signals across your entire sending history - a single campaign doesn't tank you, but a pattern of low engagement across weeks and months will
As one deliverability expert put it: "Strategies will need to earn engagement beyond AI summaries to earn the signals needed to stay in the inbox." Getting delivered isn't enough anymore. Your email needs to earn attention after it arrives, and the AI is watching what happens next.
Your Preheader Isn't Yours Anymore
For years, the playbook was simple: strong subject line, compelling preheader, hope for the open. That's breaking down.
On Apple devices, the preheader is literally being overwritten. On Gmail, the preheader still shows, but AI summaries are appearing post-open and in search results. The trend is clear: AI is increasingly deciding what text represents your email before (or instead of) the text you chose.
So what do you actually do about it?
This is where some old school tactics are actually what's working right now. We're talking about things like optimizing every image with relevant, explanatory alt text so AI actually understands what's in your email. Not "hero-image.jpg." Not "banner." Real descriptive text that tells the AI (and screen readers, and accessibility tools) what the image is about.
Structure your emails so the most important information is in real text, not buried in images. AI can't parse an image-only email. If your entire message is a designed graphic with a "Shop Now" button, Apple Intelligence has almost nothing to summarize. The AI summary might say something like "promotional email from [brand]" while the competitor who wrote actual text gets a summary like "20% off summer skincare ends tonight, free shipping over $50."
Think of your email's first 2-3 sentences as the new preheader. That's what the AI is pulling from. Front-load the value proposition, the offer, the reason this email exists. Be specific and be clear. "We just restocked the sold-out collagen peptides and added a new flavor" is going to generate a much better AI summary than "Hey! We've got exciting news for you."
Send Frequency Is Now Visible
Gmail's Manage Subscriptions feature doesn't just let people unsubscribe. It shows them exactly how many emails each brand sent in the last few weeks, ranked from most to least.
We've always been proponents of fewer, higher-quality, more relevant emails. Not fewer emails from the account necessarily, but fewer emails per subscriber. The difference matters.
If you're sending more initiatives and concepts, but each one goes to a specific segment of your list so no individual customer is getting bombarded, that's the right approach. Your account might send 15 campaigns a week, but any given subscriber only gets 2-3 because each one is targeted to people who actually care about that topic.
The brands that are going to get burned by Manage Subscriptions are the ones sending the same promotional blast to their entire list 5+ times a week. Those subscribers now have a clean, Google-designed interface that shows them exactly how often you're emailing, and a one-tap exit.
The fix isn't to stop emailing. It's to make every email you send worth receiving. Which leads to probably the most important shift happening right now.
Email Is a Content Calendar, Not a Promotional Calendar
This is the piece most DTC brands still get wrong, and it matters more now than ever because AI is essentially grading your content.
Too often, brands assume nicely designed emails full of product shots and discount codes are what ecommerce customers want. But not exactly. Customers want a connection. They want to feel understood. They want something worth reading.
We tested this with a supplement brand. We ran a plain text newsletter against a fully designed version, same content, same audience segment. The plain text version crushed it. Higher open rate, higher click rate, and more revenue per recipient.
Why? Because it felt like an actual message from a person. Not a marketing campaign. And in a world where AI is filtering emails based on engagement signals, that kind of genuine interaction is exactly what keeps you in the inbox.
Another example: we wrote a deep, educational article about peptides in supplements for a client because the topic was trending and their customers were actively searching for that information. Then we did an email newsletter about it. Not "BUY OUR PEPTIDE PRODUCT 20% OFF" but real content that answered the questions their customers were asking. It performed incredibly well.
This is the shift. Email isn't just a promotional calendar. It's a content calendar. All these things are vital now - you have to think about the bottom of the funnel as more than just pretty emails with discounts.
The brands that treat email like a relationship, not a billboard, are the ones generating the engagement signals that AI filters reward.
The Plain Text Renaissance
This might sound counterintuitive, but plain text emails are having a moment. And there's a real reason for it that goes beyond nostalgia.
Plain text emails:
- Are easier for AI to parse and summarize - there's no image-heavy layout confusing the model, just clear text that generates accurate, helpful AI summaries
- Bypass a lot of spam triggers - heavy HTML, multiple image embeds, and complex layouts are all signals that filters evaluate. Plain text avoids most of them.
- Feel personal - they look like a message from a friend or colleague, not a mass marketing campaign. That drives replies, forwards, and longer read times, all of which are positive engagement signals.
- Perform in dark mode - no broken layouts, no invisible text on dark backgrounds, no design compromises
We're not saying ditch designed emails entirely. Product launches, visual storytelling, seasonal campaigns - those still benefit from good design. But if 100% of your emails are polished HTML templates, you're missing an opportunity.
Mix in plain text. Try a weekly founder letter. Send a genuine question to a segment of your list. Write something educational that doesn't ask for a purchase. You'll be surprised at how it performs, and you might be even more surprised when it drives sales without trying to.
Open Rates Were Already Broken. Now They're Useless.
We've written about this before in our open rate benchmarks breakdown, but it's worth revisiting in this context because the problem has gotten worse.
Apple's Mail Privacy Protection (MPP), which launched back in 2021, already inflated open rates by pre-loading tracking pixels. That alone made open rates unreliable for Apple Mail users, which is roughly over 50% of email opens depending on your audience.
Now add AI summaries on top of that. If Apple Intelligence gives someone a useful summary and they decide they don't need to open, you lost the open but the subscriber still "saw" your email. Is that a failure? Depends on what you're measuring.
And here's the thing: is it really that big of a deal? Let's say someone used to open every newsletter you sent but never bought anything. They were just reading the content. Now the AI gives them a summary and they skip the open. The outcome is the same - they weren't going to buy either way. At least with the AI summary, your brand name is still in front of them.
The vanity metric obsession has to stop. Open rates aren't dollars in the bank.
What should you measure instead?
- Revenue per recipient - how much money does each email generate per person it's sent to?
- Click rate - still somewhat reliable, though not perfect
- Email-attributed revenue as a % of total - the MER-style view of whether your email program is actually moving the business
- A/B test performance - compare similar email versions against each other rather than chasing absolute open rates
- List health indicators - unsubscribe rate trends, spam complaint rates, deliverability metrics
If the business and brand are growing, and email is contributing its share of revenue, the open rate number doesn't matter nearly as much as most people think.
Technical Setup Still Matters (It's Just Not the Differentiator Anymore)
Quick reality check: SPF, DKIM, DMARC, branded sending domains, list-unsubscribe headers... all of this still matters. A lot.
It's "easy" in the sense that it's well-documented and straightforward to implement. But if it's NOT set up, it's a critical issue. Same as how not blocking your website via robots.txt is "easy," but if it IS blocked, you lose all organic traffic. Technical email authentication is the foundation everything else sits on.
Yahoo and Microsoft are enforcing authentication requirements more aggressively for bulk senders in 2026. If you haven't set up DMARC at a minimum, you're already behind.
But here's where the game has shifted: technical setup is table stakes. Everyone serious about email has it. The differentiator now is what happens above the technical layer. Content quality. Engagement patterns. Subject line authenticity.
On that note, let's talk about subject lines. Start avoiding:
- Excessive emoji usage - a single relevant emoji is fine, but strings of 🔥💥🚨 scream "marketing email"
- ALL CAPS words - "HUGE SALE" and "DON'T MISS THIS" trigger both spam filters and human skepticism
- Excessive exclamation points - "You Won't Believe This!!!" is exactly the kind of thing AI filters flag
- Vague curiosity bait - "You need to see this..." or "We have big news" gives the AI nothing useful to work with and teaches subscribers to expect disappointment
Write subject lines like you're texting a colleague about something genuinely useful. "We restocked the peptides, added a new flavor" beats "🚨 THEY'RE BACK 🚨" every time in the new AI-mediated inbox.
What to Do This Week
If this all sounds overwhelming, start with these concrete steps. You can do all of them this week.
1. Send a Plain Text Email
Just try it. Write a human, authentic, plain text email to your list about anything. Educational content, a question, a behind-the-scenes story, a genuine update. Don't put a discount code in it. Just talk to your customers like people and see what happens. We consistently see plain text outperform designed emails for engagement, and the AI signals they generate are strong.
2. Audit Your Alt Text
Go through your last 10 email campaigns. How many images have real, descriptive alt text? If the answer is "none" or "just the product name," you have a quick win sitting right there. Apple Intelligence uses your email content (including alt text) to generate its summaries. Make sure every image has context the AI can work with.
3. Check Your Send Frequency Per Subscriber
Not per account, per subscriber. Pull up your sending data and look at how many emails your average subscriber actually receives per week. If it's more than 3-4, consider whether each one is really earning its place. Gmail's Manage Subscriptions feature is going to make this number very visible to your customers.
4. Front-Load Your Emails
For your next campaign, put the most important information in the first two sentences. Not after a greeting, not after a brand introduction. Lead with the value. This is what Apple Intelligence will pull for its summary, and it's what Gmail's AI will use to evaluate relevance.
5. Review Your Subject Line Patterns
Look at your last 20 subject lines. How many of them use excessive emojis, ALL CAPS, or exclamation points? How many read like a text from a friend versus a marketing blast? Clean this up and you'll see both deliverability and engagement improve.
The Bigger Picture
All of this is part of the same trend we've been writing about across rising acquisition costs, platform changes filtering your messages, and AI reshaping how customers discover and interact with brands. The inbox is becoming an AI-mediated space, just like search results and social feeds before it.
The brands that will win aren't the ones gaming the algorithm. They're the ones creating genuinely valuable content that both humans and AI recognize as worth surfacing. That's always been good email marketing. The difference now is that the penalty for lazy, high-volume, low-value sending is more immediate and more visible than it's ever been.
Your email list is still the most valuable owned asset your brand has. That hasn't changed. What's changed is how much the quality of what you send matters relative to the quantity. The AI layer in the inbox is essentially enforcing what good email marketers have been saying for years: send better emails, less often, to people who actually want them.
If you're doing that already, these changes are actually good news for you. Your competitors who blast their whole list with the same promo email every day are about to have a much harder time.
Frequently Asked Questions
What is Apple Intelligence email summarization?
Apple Intelligence is Apple's on-device AI that automatically generates short summaries of emails. On supported devices (iPhone 15/16, M-series Macs running iOS 18+), these summaries appear directly in the inbox, replacing the preheader text you wrote. The summaries are based on the actual content of your email, not your subject line or preheader, and there's no way for senders to control or override them.
Does Gmail also replace preheaders with AI summaries?
No. As of mid-2026, Gmail's AI summaries are post-open only. They appear after someone opens your email (via the Gemini sidebar) and in search results. Gmail still shows your subject line and preheader in the inbox as you wrote them. However, Gmail's "Manage subscriptions" feature does surface your sending frequency, which can indirectly affect engagement.
How does Gmail's "Manage subscriptions" feature work?
Launched in July 2025, it's a centralized view inside Gmail where users can see all brands emailing them, sorted by who sends the most emails. Each brand has a one-click unsubscribe button next to it. Users access it from Gmail's navigation menu. The feature is live on web, Android, and iOS.
Are plain text emails really better for deliverability?
Not universally, but they have real advantages. Plain text emails avoid the spam triggers associated with heavy HTML (complex layouts, excessive images, image-to-text ratio issues). They generate accurate AI summaries because there's clear text for AI to parse. And they tend to feel more personal, which drives the engagement signals (replies, forwards, longer read times) that mailbox providers reward.
Is the open rate metric completely useless now?
Pretty close. Between Apple Mail Privacy Protection (which inflates opens by pre-loading tracking pixels) and AI summaries (which let people "see" your email without opening), open rates are unreliable for roughly half your list or more. Use open rates for A/B testing similar email versions against each other, but don't use them as a primary KPI. Focus on click rates, revenue per recipient, and email-attributed revenue as a percentage of total.
What engagement signals do mailbox providers actually care about?
The big ones in 2026: opens (still counted, just unreliable as a metric for you), clicks, replies, forwards, time spent reading, moving an email out of spam, starring or marking important. Negative signals include spam complaints, deleting without reading, and long-term inactivity. The pattern over weeks and months matters more than any single email's performance.
How do I optimize my emails for AI summarization?
Put the most important information in real text (not just images) in the first 2-3 sentences. Use descriptive alt text on every image. Keep one clear topic per email rather than cramming multiple offers. Use clear headings and short paragraphs. Write in a direct, conversational tone. Think of "summarizability" as the new above-the-fold.
Sources
- Klaviyo: AI Is Reading Your Emails First (May 2026)
- Stripo: AI Summaries in Email Clients
- Google Blog: Manage Subscriptions in Gmail (July 2025)
- Braze: How iOS Changes Are Reshaping Email Marketing
- Apple Support: Use Apple Intelligence in Mail
- emfluence: The Ultimate Guide to Email Deliverability in 2026
- UseBouncer: Email Deliverability Trends 2026
- Adobe: Gmail's Manage Subscriptions and What It Means for Marketers
- Monocle: Ecommerce Email Marketing Trends for 2026
