Published November 18, 2026

AI makes email easier to build. Making it worth reading is the hard part.

AI can now assemble a finished newsletter from your design system in minutes. That makes production easier, but deciding what's worth sending is still down to you.

Supported by Sponsor logo
On this page · 9 min read

Last week I had all the content and copy for Curated Inspiration ready to go, but I hadn’t had time to put the newsletter together before leaving for the office. Traffic was particularly bad that morning, and as I crawled across the Golden Gate Bridge, I decided to try an experiment to see how much of the newsletter production I could get done just by talking to Claude during my commute.

Surprisingly, quite a lot.

I pointed Claude to where I’d stored the content and copy, and asked it to put everything into our newsletter template in Figma using the Figma MCP. It added the links to the articles and featured emails, wrote alt text for the images, and assembled the newsletter using our existing design components.

When I got to the office, the newsletter was sitting in Figma waiting for me to review.

I’d estimate Claude had completed about 80% of the production work. I still spent time making edits in Figma and checking everything before it went out, but most of the work involved in turning my content into a finished newsletter had been done, including the tedious bits like adding links, formatting everything and writing alt text.

I’ve been using AI for all sorts of marketing work over the last year, but this was one of those moments where it really struck me how much the way we build emails is going to change.


Why it worked so well

There’s a reason Claude was able to put the newsletter together without much direction from me, and it’s not because it’s particularly good at designing emails from scratch.

We’ve been building Curated Inspiration in Figma for a while now, using the same design system every week. The format is fairly consistent, with a featured article, three interesting emails from the Email Love library, and a collection of things we’ve been reading.

The typography, spacing, image treatments and layout are already defined, and we have some basic rules about how everything works. Images need links and alt text, the featured emails have to be recent sends, and the whole thing needs to follow our existing design.

I’d also already selected the content and written the copy. Claude wasn’t having to figure out what our readers might find interesting, decide which emails to feature or come up with something original to say about them. I’d already done that work. It was taking everything I’d prepared and assembling it using components we’d already built.

I think that’s an important part of making AI useful for email production. If you give it a blank page and ask it to create something great, you’re leaving an awful lot of decisions up to a tool that doesn’t necessarily know your brand, your audience or what you’re trying to achieve.

But if you’ve already done that work, AI can take care of a surprising amount of what comes next.


Making email easier doesn’t automatically make it better

For years, building marketing emails has involved far more work than it probably should. A designer creates something in Figma, someone else rebuilds it in the ESP, a developer gets involved when something doesn’t render correctly, and then everyone goes through another round of reviews to make sure the finished email matches the original design.

We’ve been working on removing that friction with Email Love, and now AI is making it possible to automate even more of the production process.

I think that’s great, because I’ve spent enough time working in email marketing to know how much effort gets wasted rebuilding things that have already been designed or repeating the same production tasks every week.

But there’s another side to this that I’ve been thinking about.

When it becomes incredibly easy to produce a professional-looking email, what happens to the quality of the content we send?

My guess is that a lot of marketing teams will use AI to produce more campaigns, more personalized variations and more automated emails. Some of that will be genuinely useful, particularly when it means customers receive more relevant information.

But I also suspect we’re going to see an awful lot more mediocre content being produced simply because it’s now so easy to make.

You can already open Claude, ask it to write an article about the latest trends in ecommerce marketing, and have something reasonably well written in under a minute. It’ll probably cover all the obvious points, explain why personalization matters, talk about AI and automation, and finish with some sensible recommendations.

There’s nothing necessarily wrong with it, but there’s a good chance you’ve already read the same article a dozen times.

Email is heading in the same direction. AI can write decent copy, follow brand guidelines and produce attractive layouts, but that doesn’t mean the finished email has anything particularly interesting to say.

And as more companies start using the same tools to produce more content, I think it’s going to become increasingly difficult to get people to pay attention.

Your subscribers don’t care whether it took three days to produce your campaign or three minutes. They just want to know whether it’s worth opening.


What actually makes an email worth reading?

This is the part I think marketers should be spending more time on, particularly if AI is going to take care of more of the production work.

Not every email needs to be particularly creative or contain some groundbreaking insight. If I signed up to hear when a product comes back in stock, an email telling me it’s available again is useful. The same goes for a genuinely good offer, a helpful product recommendation or information I’ve specifically asked to receive.

But when you’re sending newsletters, educational content or campaigns intended to build a relationship with your audience, I think you need to offer something more than a well-written summary of information people can find anywhere.

Curated Inspiration is a good example of this.

We have hundreds of thousands of emails in the Email Love library, and every week we pick three to feature in the newsletter. Finding three attractive emails isn’t particularly difficult, and AI can certainly help with that.

The harder part is deciding which ones are actually worth sharing.

Maybe a brand has done something interesting with its layout, found a clever way to personalize an email, or taken a different approach to promoting a product. Whatever it is, I want to be able to explain why I think it’s good and what another email marketer might learn from it.

That’s the reason someone subscribes to Curated Inspiration in the first place. They’re not just looking for three screenshots of emails. They’re interested in discovering good work and understanding what makes it good.

And this is where I think a lot of AI-generated content falls down. The writing might be perfectly fine, but if nobody has put much thought into what they’re saying, you end up with something that sounds like every other email or article on the same subject.

In my bridge experiment, I’d already selected the content, written the copy and decided why each story was interesting. Claude didn’t need to make those decisions. It just needed to turn everything I’d prepared into a finished email.

If I’d asked it to come up with the entire newsletter, including selecting the emails and writing about why they were interesting, I’m not convinced the result would have been nearly as good.

That’s not to say AI can’t help with research or come up with useful suggestions. I use it for both all the time. But I still want to make the decisions about what we publish, what we recommend and what we think our readers will find valuable.


What we should do with the time AI saves us

Most of the discussion about AI in marketing seems to focus on productivity. How much faster can we create a campaign? How many tasks can we automate? How much more content can we produce with the same team?

Those are all reasonable questions, but I wonder if we’re measuring the wrong thing.

If AI saves me a few hours putting together Curated Inspiration, I could use that time to create another newsletter. Or I could spend more time going through the Email Love library, finding examples I might otherwise have missed, and thinking about what our readers would actually find interesting.

I’d much rather do the second.

I also think there’s an opportunity to improve the quality of the emails themselves. If we’re spending less time assembling templates, rewriting HTML and fixing the same rendering problems, there’s more time to test our work properly, improve accessibility, try different creative approaches and make sure we’re sending something we’re happy with.

Of course, you don’t get those improvements automatically just because you’re using AI. You still need to have a good idea of what you’re trying to achieve and some standards for what good looks like.

That’s one of the reasons we’ve put so much effort into design systems at Email Love.

Rather than asking AI to invent a new layout and write HTML every time you build an email, we want it to work with the components and designs your team already uses. Our Figma plugin takes care of converting those designs into production-ready HTML, and Email Love Builder uses the same rendering and export engine for teams that prefer working in a drag-and-drop editor.

We’ve also released free Claude skills to help AI build emails and design systems in Figma, and the Email Love MCP gives it access to real emails from our library for research and inspiration.

The idea is to take care of as much of the repetitive production work as possible, without having to compromise on the quality of the finished email.

But that’s only half of the problem. We can make it much easier to build an email that looks great and works across different inboxes, but whether the content is worth reading is still down to the people creating it.


I’d rather send one better newsletter than five average ones

I came away from the bridge experiment feeling pretty optimistic about where all of this is heading.

A year or two ago, the idea that I could talk to Claude on the way to work and arrive to find my newsletter already assembled in Figma would have sounded a bit ridiculous. Now it’s something I can actually do, and I expect it’s only going to get easier.

What I don’t want is to use that efficiency to fill people’s inboxes with even more average content.

If anything, I think the opposite should happen. When we don’t have to spend so much time building emails, we should be able to put more effort into making them genuinely useful, interesting and relevant to the people receiving them.

I’d rather send one really good newsletter every week than use AI to churn out five average ones, and I suspect most subscribers would prefer that too.

We’re getting to a point where almost anyone can produce a professional-looking email quickly and cheaply. That’s a good thing, but it also means that simply being able to produce a good-looking email isn’t going to set you apart anymore.

The more interesting question is what you’re going to put in it, and whether anyone is going to want to read it.

Much love,
Andy

About the author
Andy King
Andy King
Co-founder, Email Love

Andy has spent two decades building email programs for brands and agencies, and now co-runs Email Love: the inspiration gallery of 628,000+ emails and the Figma plugin that lets teams design, build, and export production-ready email without leaving their design file. He writes about email design systems, production workflow, and where the industry is heading.

Say more, more powerfully

Customer.io delivers more impact from every message–unifying data, messaging, and AI into one platform so every customer interaction gets smarter over time. We help teams build personalized campaigns across email, SMS, push, in-app, and webhooks, using their own first-party data to reach the right people with the right message at the right moment. AI is built into the platform to remove the grunt work, letting marketers focus on strategy, not setup. Over 9,000 brands trust us to power their messaging.