The New Inbox AI Middleman and What it Means for Marketers

By Emily Airey, Marketing Executive at Force24

 

As marketers, we have spent years trying to win the email open. We’ve obsessed over subject lines, tested send times, refined preview text and worked hard to earn that one small signal that says someone chose to have at least a look. But now, I think the challenge has changed. Increasingly, the question is not just whether an email gets opened, it’s instead now whether the message survives being interpreted, summarised and filtered before the reader ever properly sees it.

 

Enter the middleman…

The DMA’s Email Council has already identified this as one of the defining shifts for email and CRM teams in 2026. Their view is that inbox behaviour is changing, AI is influencing how emails are read, and marketers need to rethink both design and measurement as a result. They specifically note that AI tools built into platforms such as Gmail and Outlook can summarise emails and surface key information without the user opening them, which means a message can be consumed without generating a traditional open. It also recommends placing critical information at the top of emails, prioritising clear subject lines and pre-headers, and using accessible text that AI systems can summarise.

For me, that changes the brief for email strategy.

For a long time, good email writing has often involved holding something back. Tease the value, create enough intrigue to encourage the click. Delay the payoff just enough to earn attention. My CEO, Adam, has always described this as giving them an appetiser, not the full meal. That approach still has its place, but it becomes far riskier when inbox tools are extracting the gist of an email before the recipient decides whether to engage further. If the summary is vague, cluttered or misleading, the message can lose value before we the marketer have had a fair chance to make our case.

I use AI most days for my own job and so assumed almost everyone had made it part of their day, however when speaking to friends in non-digital sectors noticed there was a lag of usage outside of meal recipes and such. However that is quickly catching up, with Ofcom stating that more than half of UK adults now use, which is up sharply from the previous year, with even higher uptake among younger adults. That matters because it shows AI-assisted interaction is becoming normal, not novel. Consumers are growing more comfortable with software interpreting information on their behalf, and that habit will not stop at search or chat interfaces. It is already affecting how people expect digital content to work.

 

So, what does this mean for email marketers in practical terms?

First, it means clarity is moving up the priority list. I do not mean bland copy. I mean clear, useful communication that makes sense quickly. If an AI summary is likely to pull from the opening copy, the subject line, the pre-header and the visible structure of the email, then those elements are doing even more work than they were before. We cannot afford to treat them as wrappers around the message. In many cases, they are now the message.

That is why I think the old distinction between “creative” and “functional” email copy is starting to break down. In an AI-mediated inbox, your most effective email may not be the smartest one. It may be the one that communicates purpose most cleanly, presents value most early, and makes the next step most obvious. My guidance would be in exactly this direction, advising marketers to keep layouts simple, use more text-based structure, ensure HTML is semantically clear, and design on the assumption that many users may never fully open the email at all.

 

Second, it means we need to think much harder about hierarchy

A lot of marketing emails I see still bury the point. You see a strong visual at the top, a brand statement, followed by a broad intro. I’m probably guilty of this from time to time as well, as it’s what always felt like a safe recipe. Then eventually, somewhere lower down, we see the practical reason why the email exists. If your key commercial message only appears halfway through, there is a good chance the inbox layer will never reflect it properly.

This is one reason I think marketers should revisit the discipline of above-the-fold messaging. This is not because everyone now needs to send shorter emails, but because everyone needs stronger prioritisation. The opening section should answer the essential questions quickly: what is this about, why does it matter to me, and what should I do next? 

 

Third, and this is a big one, it means open rates matter even less than they did before

Most experienced email marketers already know opens have become messy. Privacy protections, automatic pre-fetching and security scanning have all chipped away at their reliability. But AI summaries add another complication. As the DMA puts it, messages can now be consumed without generating a traditional open, and image carousels or preview experiences can reduce recorded opens further. I agree  that is the right conclusion.

If AI is changing how messages are previewed, filtered and partially consumed, then performance measurement has to move closer to commercial reality. Did the email generate a meaningful visit? Did it move someone deeper into a journey? Did it support enquiry, purchase, booking, retention or repeat action? Those are far more useful questions than whether a pixel fired. But this does not mean opens become completely useless. It means they become even less suitable as a headline KPI. 

The more the inbox starts doing some of the reading, the more marketers need to judge email on its ability to drive outcomes.This is also where the quality of the underlying email matters. The DMA’s Marketer Email Tracker 2026 saysthat AI-powered inboxes now influence how messages are prioritised and summarised, and that the opportunity lies in using data to create relevant, timely and lasting customer relationships. It also points to segmentation, lifecycle tracking and customer lifetime value measurement as established strengths across the market. That is an encouraging observation for me, because it suggests the future of email will not be won through higher volume, but instead better relevance.

When our inboxes become better at identifying what looks useful, generic campaigns become even easier to ignore. Emails that are weakly segmented, overly broad or slow to get to the point are likely to struggle more, not less. By contrast, emails that are timely, behaviour-led and grounded in clear customer context should benefit. That is why I do not see this as bad news for email. In many ways, AI in the inbox is forcing a return to the fundamentals good marketers should value anyway: 

 

What to do now

From where I sit, the response is fairly straightforward. Here are 6 things you can do to start aligning to what comes next:

1. Start writing subject lines and pre-headers that communicate meaning, not just intrigue. 
2. Put the most important information earlier.
3. Reduce over-reliance on image-only storytelling. 
4. Make email copy easier for both humans and machines to interpret. 
5. Review your metrics, KPIs or reporting frameworks and make sure they don’t lean too heavily on opens, clicks. Instead align them to commercial goals. 
6. Most importantly, but the hardest – if they aren’t already, build your workflows and journeys around behaviour and relevance.
   

If you found this article helpful, we’ve just launched a new Inbox Intelligence series that you might like. This is a video series l designed to help marketers master the basics and beyond. We release new episodes every month, with every episode designed around realistic improvements you can introduce immediately. You can watch all episodes here.

Back to FWD: Thinking