A practical look at what and how AI Previews being used by the major mailbox providers have done to impact deliverability.  And what you need to do to make sure your campaigns actually reach the inbox.

Short Summary of What You Will Learn: We wrote this article to help you with email deliverability. We explain the impact of AI Previews and how this impacts how to build emails.  We explain how/why its important to lead with real text and a strong subject line. We recap how to pass SPF, DKIM, and DMARC. And finally we discuss other ways to keep your spam-complaint rate under 0.10% to 0.30% and the benefits of data hygiene and well segmented lists.

We have also summarized the action items from our learnings in a full checklist. You can access this checklist in two ways. You can download a PDF for easy access at any time. Or you can reference online in this supporting blog post Checklist for AI Era Inbox.

Background

We just attended the Salesforce Connections conference in Chicago.  At the conference nearly every session addressed new AI capabilities across the Salesforce and Marketing Cloud family (including Sales Cloud, Account Engagement, Slack and  Marketing Cloud Next). There is a great deal to be excited about especially what will be possible with native AI tools in Marketing Cloud Next.

But what stood out as immediately useful/actionable is this: the major mailbox providers have used AI to quietly change what it takes to consistently land your emails in your target’s inbox. There are three main issues that every team needs to immediately address to maintain and ensure successful delivery:

• AI previews and how they change the way you need to build an email
• Authentication and sender reputation
• List quality

As positive news, even with all the changes occurring with AI , Salesforce data continues to confirm that email marketing and lead nurturing still produces one of the best and most cost effective ROIs for B2B companies. In this short article we will explain what has changed and what needs renewed strategic attention/focus to ensure email deliverability.  We will also explain why it matters so you can make the right strategic decisions for your organization.

AI Previews: Build for the Machine That Reads First

Before a single human sees your email, an AI may have already read it and decided what it’s about. Apple Mail uses Apple Intelligence to write a short summary from the first few hundred words of your email and shows it in the inbox in place of your preheader. As an email marketer you can’t override it or turn it off. Gmail offers Gemini-generated summaries as well. These tools also help decide what’s most important inside the message, which means structure, content hierarchy, images, alt text, and the placement of your key information all matter more than they used to.

One useful detail from the sessions: Apple’s summary is built from your live body text and largely ignores your subject line, your preheader, and any text inside image alt tags. So how you assemble the top of your email now shapes the preview your reader sees before they decide to open. The encouraging part is that the same choices that help the AI also make your email more accessible to screen-reader users and easier for busy people to skim. Five practices make the difference. For more, see Salesforce’s tips for navigating AI email summaries and Litmus’s breakdown of what AI summaries mean for marketers.

1. Put the key message in the first few sentences

Lead with the single most important takeaway, such as the offer, the value, the reason to act, in your opening lines, not after a greeting and a paragraph of warm-up.

Why this matters: Apple Intelligence builds its inbox summary from the first few hundred words of your body copy. If those opening lines are “Hi there, hope you’re having a great week” or housekeeping text, that’s what the AI summarizes and that becomes the preview your subscriber judges before opening. Front-load the substance and the AI has something worth showing.

2. Use real text rather than text embedded in images

Keep your core message in live, selectable text. Don’t bake your headline, offer, or key copy into a graphic.

Why this matters: AI can only summarize text it can actually read it. Text trapped inside an image is invisible to it, just as it is to screen readers and, often, to spam filters. Image-only emails give the summary engine almost nothing to work with, and testing shows summary quality drops sharply for them. Don’t lean on alt text as a safety net, either: AI summaries frequently ignore it. Live HTML text is what carries your message.  Alt text remains critical for accessibility, but today’s AI inbox summaries appear to rely primarily on the visible email content rather than the alt text attached to images.

3. Use a clear preheader

Write a specific, benefit-driven preheader that extends your subject line, however, your subject line needs to stand alone in case some never see your preheader.  Do not use “View this email in your browser” or a string of empty filler.

Why this matters: The preheader is still the preview line in many inboxes and for every recipient who isn’t seeing an AI summary. A clear preheader earns more opens, and where AI summaries do appear, a focused top-of-email keeps your signal consistent. It’s a small field that quietly shapes a first impression.

4. Don’t stuff the top of the email with legal text, navigation, or “view online” links

Move boilerplate nav menus, legal disclaimers, “view in browser” out of the opening. Lead with the message; let the housekeeping sit lower or in the footer.

Why this matters: AI reads top-down and weights what comes first, so if the top of your email is a navigation bar and a disclaimer, that’s what can end up in the summary and that will be reader’s first impression. Clearing the top lets both the AI and your reader get to the point.

5. Use semantic structure: headings, bullets, and concise paragraphs

Organize the email with real headings, short paragraphs, and scannable lists rather than one dense block of copy.

Why this matters: Clear structure helps the AI understand your hierarchy and identify what’s most important, and it helps human readers skim and act. It’s also the foundation of accessibility. Screen-reader users navigate by headings so good structure serves the machine, the assistive technology, and the busy reader all at once.

Authentication and Sender Reputation: Now a Gate, Not a Best Practice

SPF, DKIM, and DMARC used to be the kind of technical hygiene that lived on someone’s “we should get to that” list. Not anymore. In plain terms, these are the records that prove an email actually came from you and wasn’t spoofed by someone else. Gmail, Yahoo, and Microsoft now treat them as baseline requirements. And now for higher-volume senders, mail that fails authentication is rejected, not just filtered to spam. A great campaign that fails authentication simply never arrives.

What these terms actually mean

SPF (Sender Policy Framework): a public list of the mail servers allowed to send email on behalf of your domain, so a receiving server can confirm your message came from an approved source. Google’s SPF overview.

DKIM (DomainKeys Identified Mail): a tamper-evident digital signature added to each message, letting the receiving server verify the email genuinely came from your domain and wasn’t altered in transit. Google’s DKIM overview.

DMARC (Domain-based Message Authentication, Reporting & Conformance): the policy that tells receiving servers what to do with mail that fails SPF or DKIM — deliver, quarantine, or reject — and sends you reports showing who is sending under your domain. Google’s DMARC overview.

Missing or misconfigured authentication is the single most common gap we see, and it’s usually invisible until deliverability quietly erodes. If you haven’t confirmed your domain is passing all three, that’s the first thing to check. See Google’s email sender guidelines and Microsoft’s requirements for high-volume senders (enforcement began May 5, 2025).

For a clear breakdown of the new bulk-sender rules and why this matters, see Validity’s guide.

List Quality Protects Everything Discussed Above

All of the above rests on list quality. Because providers reward engaged senders, consistently mailing inactive or invalid contacts actively drags down deliverability for your entire list — your engaged subscribers included. List hygiene has moved from housekeeping to strategy requirement. Regularly removing dead addresses, suppressing chronically unengaged contacts, and segmenting by real activity isn’t about a tidier database for its own sake; it’s about protecting the inbox placement of the people who do want to hear from you. This is also where clean, well-organized data in your CRM and marketing platform pays off directly. Data governance and data hygiene are becoming more important.

The 60-Second Deliverability Checklist

If you only do four things, do these. The rest of this guide explains the why behind each one.

1. Confirm your domain passes SPF, DKIM, and DMARC (this is now a hard requirement, not a nice-to-have).
2. Build every email with real text and a clear, standalone subject line.
3. Write a deliberate preheader — don’t let the platform invent one.
4. Mail engaged contacts; suppress and remove the rest on a regular schedule.

For a longer more comprehensive checklist download our PDF here or visit our blog post  Or you can reference online in supporting blog post Checklist for AI Era Inbox.

The Bottom Line

AI is changing how people consume email, but deliverability is still won or lost through authentication, sender reputation, engagement, and list quality.

The encouraging part: most teams already own the tools to do all of this, and it’s usually a matter of operationalizing how you build emails, your authentication and DMARC monitoring, and your list hygiene, rather than buying anything new. If you’d like a second set of eyes on where your email program stands today, that’s exactly the kind of behind-the-scenes work we help our clients put in place at Gabriel Sales. We’re always happy to talk it through and staff the resources to help you if necessary.

About the Author 

Carol Springer is a Founder at Gabriel Sales.  She has been working at the inflection point of Marketing an Sales Operations implementing an optimizing Salesforce and Marketing Automation for the past 15 years.  Carol has multiple Salesforce certifications leads the Denver Colorado  Salesforce Marketers User Group and is a frequent Salesforce Speaker.  Prior to focusing on Sales and Marketing Ops Carol was an award winning enterprise sales executive. 

Sources & further reading

• Salesforce — 5 Tips to Help Marketers Navigate AI Email Summaries
• Litmus — AI-Generated Email Summaries: What Marketers Need to Know
• Google — Email sender guidelines
• Google — Set up SPF
• Google — Set up DKIM
• Google — Set up DMARC
• Microsoft — Outlook requirements for high-volume senders
• Validity — Microsoft’s new bulk email rules, explained