I've got a new theory of information I'm going to share with you here, and since it's so new, I'd love to know what you think.

How many newsletters do you subscribe to with potentially useful information? For me, I probably get 5-10 daily, another 10-20 weekly, and then a smattering that come monthly or whenever the heck someone feels like sending it. A few of them are even published by outlets other than Marketecture. Maybe Ari will forgive me.

Now, let's think about how useful that information is in such newsletters. You could sort all that information into four categories:

Theoretically, most information in any given newsletter is in that first “ignore” bucket. And a bucket's a good way to think about it. We're throwing a bucket into the water and seeing if any yummy fish get in there. We have our own tolerance levels. If you know me, for instance, you might forgive a long stretch of empty buckets when you read this. For most readers who I haven't met, you'll need to catch something way sooner and more often, or you're not coming back.

Some information is timely. If you're getting emails about events or sales, those emails have a short shelf-life. That's true for breaking news. If you missed who won March Madness, by now you probably don't care (and you probably didn't find out in a newsletter, anyway). 

If you're learning about new feature updates for products you use, it's likely that you won't have to use that information immediately. But the sooner you get to it, the better.

For more educational content, however, the shelf life could last years. 

As an example, Tom Orbach writes the terrific Marketing Ideas newsletter. I don't know how he sources so many good ideas every week. That's what makes his newsletters invaluable. 

But it's often hard to do anything with them the day I read them. They're inspirational, sure. But I might not be running a trade show that month, or I might not be in a position to buy cheap billboards as a marketing stunt. When I need inspiration for my next marketing stunt, though, I'll want to inject his ideas straight into my veins. 

I've had a habit of saving some of my favorite such newsletter editions as PDFs, storing them in a folder on my desktop. In theory, I can take these and create some kind of Custom GPT or NotebookLM and then reference these editions in an interactive format; I've done this at times, and it's useful, and I have other favorite sources beyond Tom.

But now, more AI tools can connect to your email directly. ChatGPT, Claude, Gemini, and other AI apps have such connectors. You can reference all the information from newsletters on-demand. And since this is a library of newsletters that you subscribe to, it's a more valuable database than the entire corpus of publicly available sales and marketing newsletters. You curated it, and whether or not you consumed the information when it was delivered, you have this on-demand, interactive library. 

Newsletters are uniquely suited for such a use case. Content is longer form but easily scannable and searchable. It's not buried in a podcast or video, and much of the best content is typically delivered in the body of the newsletter, rather than through links. Once you archive this newsletter (you'd never trash it, right?), it might come up in a search 20 years from now. You'll be a hologram. Greenland, in a surprise twist, will have purchased the U.S. But this newsletter will still be in your inbox, and translated into our national language of Kalaallisut.

Hence, the Theory of Stored Newsletter Value states that the value of non-time-sensitive newsletter information increases over time, and that value is currently on a rapidly escalating curve (which will level off) as AI-powered technologies improve. 

To better examine this theory, let's look at three different potential scenarios:

1) Nothing changes. The value of newsletters is just the same as it was five years ago, and it'll probably be the same five years from now. 

2) Newsletters' stored value skyrockets. Newsletters that share more content within the body of the email become more valuable over time, and they will benefit from lower unsubscribe rates. Some immediate value for advertisers diminishes, as readers pay less attention to some newsletters in hopes that they'll be more useful later. Long-term latency effects expand. Advertorial-based models and text-based in-stream ads become more valuable since they become part of the content, which AI then processes as useful information. Banner ads over time get less interaction.

3) AIOD: AI-On-Demand kills the newsletter business entirely. Who needs newsletters when AI curates and generates all the content you need? Newsletter subscriptions plummet. AI agents proactively surface most of the time-sensitive info anyone needs.

Scenario 2 is more of that Newtonian equal and opposite reaction to Scenario 3. AI proliferates but gives newsletters a new life. The old and the new flourish together.

As I was writing this, Google sent me a note with this product update: "Your Google AI Pro plan has been upgraded to include 5 TB of storage at no extra cost to you." Google may be facilitating more content creation through generative AI, but it's also making it possible to reap more benefits from stored value.

The biggest question is whether we'll still have the attention spans to read newsletters ourselves, or if we'll be reliant on AI to read them for us.

— David Berkowitz, Chief Community Officer, Marketecture Media

Who: Agency Principals, Retail CMOs, E-commerce Directors

What: A new report reveals that 67% of retail executives expect to have full AI-driven personalization capabilities by the end of the year, with three-quarters planning to reduce reliance on external agencies. Retailers are increasingly building in-house Creative Operating Systems to generate personalized product descriptions and visuals at a volume human teams can't match.

Why it matters: Agencies that provide execution are being replaced by in-house bots; to survive, agencies must pivot to high-level strategy and AI-stack management.

(Deloitte)

Who: Retail CMOs, E-commerce Teams, Brand Innovation Leaders, D2C Marketers

What: Macy's quietly launched an AI shopping assistant called Ask Macy's, and the early data is striking: Customers who used it spent 4.75 times more per visit than those who didn't. Details on the rollout remain limited, but the result reinforces a growing pattern. AI-assisted shopping drives significantly higher conversion and basket size than standard browsing.

Why it matters: A spending lift like this could accelerate AI shopping assistant deployments across retail, and it raises the bar for brands that haven't moved yet.

(CX Dive)

Who: CMOs, Growth Marketers, Brand Safety Leads, Healthcare Marketers, Agency Strategists

What: A buzzy telehealth startup was celebrated as a one-person, billion-dollar company powered by AI. But scrutiny quickly revealed questionable AI-generated marketing, fake personas, and regulatory warnings — raising doubts about whether the growth story reflects real innovation or risky shortcuts.

Why it matters: Not every AI miracle is a blueprint. Marketers need to separate scalable AI use from hype, especially when growth tactics cross ethical or legal lines.

(Forrester)

Liking AI Brief? The Refresh is a weekly recap of all the most important advertising news, written by AdTechGod.

Thanks for reading this week’s AI BRIEF by Marketecture. 

Marketecture Media is committed to bringing you the latest news, opinions, and community updates. 

If you are interested in advertising on this newsletter, our podcasts, The Advertising Forum, or in our Slack community, you can contact us here.