Workweek Newsletter {beacon}

8 AI changes Marketers need to know

There was a LOT of AI news this month. I pulled out the stuff that actually changes how we market. Here’s what made the cut.

Daniel Murray

Sep 13th, 2026

{cta_url_read_in_browser = community_base_url + "/library/" + article_id + "?utm_source=newsletter&utm_medium=email&utm_campaign=" + edition_slug + "&utm_content=read_in_browser"}{cta_url_read_in_app = community_base_url + "/library/" + article_id + "?utm_source=newsletter&utm_medium=email&utm_campaign=" + edition_slug + "&utm_content=read_in_app"}{cta_url_join_conversation = community_base_url + "/library/" + article_id + "?utm_source=newsletter&utm_medium=email&utm_campaign=" + edition_slug + "&utm_content=join_conversation" + "#comments"} {if profile.vars.member_status == "lead" || profile.vars.member_status == "unfit"} Read in browser {else}{if profile.vars.member_status == "fit"} Read in Marketing Land {else}{if profile.vars.member_status == "member"} Join the convo {else} Read in browser {/if}{/if}{/if}

In partnership with

Hey Marketing Bestie,

Will you be in Boston this week at Unbound? There's a lobster roll (or 4) with my name on them and I am really looking forward to the trip.

If you see me, please shout. I'll be the guy in at least double, possibly triple, The Marketing Millennials merch. At LEAST a TMM hat.

Maybe I'll bring you 1?

Was this email forwarded to you?

Subscribe Here

AI Marketing News This Month

1️⃣. Your Email Now Has a Second Reader...And It's Not Human

For years, your email had 1 audience: the person who opened it.

Not anymore. Not in any way.

Gmail is rolling out AI Overviews that summarize inbox content and surface answers without a human ever clicking open.

Which means your email now has 2 readers.

1 is a person. The other is a machine deciding whether your message is even worth surfacing.

Bury your offer in a clever intro or a giant perfectly designed image, and the robot skips it entirely.

Put the price, the deadline, the CTA in plain text. Make it SUPER obvious. That's the only way both readers see it. I've told you this before but it really matters more than ever so starting this list there. Thank you!

2️⃣. LinkedIn Just Told Marketers How To Write For The AI Reading It

For years, a LinkedIn post had 1 job: perform in the feed for the next 48 hours.

Then it was gone.

LinkedIn's own Marketing Solutions team published guidance on August 11th of this year telling Marketers to write differently - clear headers, direct answers, recurring topic clusters, a consistent point of view. Because the post isn't just competing for attention anymore. It's competing to become source material.

The post you write for 20,000 followers today can end up answering a stranger's question in ChatGPT 6 months from now. Which is great as we want content to have a long shelf life.

So stop posting random thought leadership.

Pick 3-4 topics you want your brand associated with, and answer specific questions about them, over and over, until the pattern is unmistakable.

That repetition isn't just brand-building anymore. It's what gets you cited. LinkedIn is the channel I know best. Listen to me on this, please.

3️⃣. The New Subscribe Button Lives Inside Google Now

For years, algorithms decided who saw what.

Google just handed some of that back...readers can now mark you as a "Preferred Source" across Search, Discover and News, and over 600,000 sources have already been selected.

Which means an audience you build can directly influence how often you get shown.

Start treating "make us a Preferred Source" the way you treat "subscribe to our newsletter."

Because now, it basically is 1.

4️⃣. Amazon Just Handed Advertisers A Direct Line Into ChatGPT

This 1 happened this week.

Amazon Ads announced an integration with OpenAI letting select U.S. advertisers extend Amazon DSP campaigns into ChatGPT ad inventory.

Amazon knows shopping behavior. ChatGPT knows what someone's actively asking about.

Those 2 things are starting to collide.

OpenAI's CFO called it basically what happens if Google and Meta had a baby.

AI assistants just became a media channel you can buy into, not just one you hope to show up in organically.

If you're in Ecommerce, treat your product feed like creative. Titles, descriptions, categories, images. My wife Ari is Chief Digital Officer at Salt & Stone and I ran to find her when I read this news. It's major for Marketers like her who are responsible for DTC and Amazon.

The cleaner your data, the easier it is for AI to know when your product belongs in the answer.

5️⃣. Google Just Turned Search Console Into A Social Analytics Tool

Your SEO team and your social team have been reading two different dashboards for years.

That's over.

Search Console can now show how your TikTok, Instagram, YouTube, and X content performs inside Google Search. Which posts get clicks, what queries drive them, Reels versus posts, Shorts versus long-form.

Add your social accounts as Search Console properties.

Then every week take your fastest-rising query and build 1 more piece of content around it.

6️⃣. 94% Of People Would Buy A Brand They've Never Heard Of

CRAZY STAT. 94%. That's everyone honestly.

You spent years building brand awareness. Building the thing everyone said mattered most.

Razorfish found that might matter less than you think: 94% of AI users said they'd consider a brand they'd NEVER heard of if the AI recommended it, and 57% actually bought from a brand they hadn't originally considered.

Being famous doesn't guarantee you the shortlist anymore. Being cited does.

So go audit what AI can actually find about you: reviews, Reddit threads, comparisons, editorial coverage, creator mentions, customer discussions. That's your real brand equity now, not your follower count.

7️⃣. Creators are becoming an AI-search strategy

Creator marketing used to have one job: exposure to a trusted audience.

Digiday reported on August 20th of this year that brands are now using creator content for something else entirely: increasing their odds of showing up in AI recommendations.

Here's why: LLMs pull from public third party content, not from a brand's own website.

Which means the creator posting about your product might matter more to an AI answer than anything you've ever published yourself.

So the brief changes. Don't just ask a creator to post about you.

Ask them to answer a real question, explain a real comparison, or lay out exactly who your product is for.

8️⃣. Your Product Page Isn't What ChatGPT Shows Anymore

For a while, nobody knew why ChatGPT picked 1 product over another. It just... did. It was a total black box.

Now OpenAI's laying it out: structured metadata, third-party reviews, pricing data. It does rely on your marketing copy OR your beautifully written headline.

ChatGPT even writes its OWN titles and labels, stuff like "Budget-friendly," built from whatever data it finds, not whatever the merchant actually wrote.

There's only 1 way I translate it: your polished product page might get rewritten before a shopper ever lays eyes on it.

So stop optimizing the page. Start optimizing the data underneath it: clean pricing, real reviews, accurate metadata. Messy data means

ChatGPT fills the gaps with its own version of your product, and you don't get a vote.

MEME OF THE WEEK

I miss this trend. Life was more simple back then.

WORKWEEK