Your weekly guide on everything AI marketing. Covering everything from AI marketing news, tips, deep dives, events, podcasts, jobs and much more. Never miss a beat with our 5-minute newsletter.
In today’s email:
Organic traffic fell 10.5%, and the visitors who stayed are worth more
X hands the ad account to an agent that can spend real money
Gradial builds brand-new ads out of parts you already own
Plus: an agency president on the shift that stopped rewarding scale, Ross Simmonds on the channel marketers still underrate, and why two thirds of shoppers are bringing AI to Christmas.
The Top 3 Stories
1. Your search traffic fell. Your best visitors did not.
Brainlabs pulled Google Analytics data across 54 clients and found organic sessions down 10.5%, from 140.1 million to 125.4 million, with 46 of the 54 declining. Sessions from AI platforms rose 163%. Visitors from ChatGPT, Copilot, Gemini and Perplexity completed key events at roughly 1.5 times the organic rate.
Why it matters: If your board still grades SEO on sessions, you are about to defund your best-converting channel. 📰 Digiday
2. X hands the ad account to the agent
X shipped a Model Context Protocol server on 24 August exposing 23 tools to AI agents, ten of which write to live, funded ad accounts. Campaigns created this way arrive paused, so a hallucinated targeting line costs nothing until something activates it. Grok and Claude Code are the named clients.
Why it matters: Your ad account’s security boundary is now whoever holds the token, not whoever has the login. 📰 PPC Land
3. Gradial builds new ads out of parts you already own
Gradial shipped capabilities that assemble entirely new on-brand assets from approved backgrounds, copy, logos and product imagery already sitting in Figma or SharePoint, rather than generating from a prompt. Chief executive Doug Tallmadge says it can now produce the 1,500 or so variations a campaign needs. It bills on marketing output rather than token use.
Why it matters: When variations cost almost nothing, your constraint becomes the quality of the assets you already own. 📰 AdExchanger
More trending AI marketing news from last week
Shopify, OpenAI and Cloudflare took WebMCP live, letting agents use sites directly.
Google added link carousels for developing topics inside AI Mode.
Omnicom moved 468 Omni engineers to Endava while keeping the platform.
Crescendo launched AI agents to run the whole customer service stack.
Together with GoHalves
Helping brands grow and scale faster through AI-driven, co-marketing matchmaking.
GoHalves is an AI matchmaker for co-marketing. A short interview learns what you do and who you reach, then finds the verified, non-competing brands whose audiences genuinely overlap with yours.
When you both accept, the campaign arrives already drafted, with tracking links so you see what it delivered. Newsletter swaps, co-webinars, podcast guest spots, guest posts. Run it as is, edit it, or counter.
That same interview is how you become a founding member. Joining is free and commits you to nothing, and founding members get 50% off Pro or Enterprise for twelve months with nothing charged until matching opens. Save your spot today:
Quote of the week
For the first time in my 30-plus years in the agency space, a major technological shift is not rewarding scale.
Trending AI tools for marketers
Here are 4 trending tools to explore this week. Quick, practical upgrades to your AI marketing arsenal:
Screenify Studio describes a demo in plain English and an AI records it for you.
Tellie follows the words you actually say and flags what you skipped.
Lenz fact-checks AI-written claims against independent sources before you publish.
LoupeKit scores how much of any web page reads as raw AI output.
TV, podcasts & streaming
Marketing Over Coffee, “Raise the mug for 9 to 5”
The episode covers running language models locally and getting better writing out of AI. The local-models thread is the one that matters for marketers, because it decides whether client data ever leaves your own machine, listen to it here.
AI training
How to record a product demo in under 30 minutes using Screenify Studio
The overview:
Screenify Studio drives a real browser from a plain-English description and hands back a staged product demo video, so a launch clip takes about half an hour.
Step-by-step:
Write the demo as one paragraph
Describe the exact path you want a viewer to see, screen by screen.
Vague briefs produce wandering takes, so name the buttons and the end state.Sign in once, by hand
Log into your product yourself and save the session to a named profile.
Almost nothing interesting happens on the logged-out page, so this step is not optional.Let the agent record the run
Point it at the URL and let it click through while you watch.
Re-run the same script after a UI change rather than re-shooting the whole thing.Stage it, do not decorate it
Drop the recording into a device frame and pick one camera move.
Good output looks like a product, not a showreel, so resist stacking effects.Export a short social cut
Trim the strongest twenty seconds and export it separately.
The full demo belongs on the page, the short cut is what earns the click.
Pro tip: Save the script alongside the feature ticket, so the demo gets re-recorded automatically every time the interface changes.
The weekly deep dive
Shoppers are already planning on automating Christmas with AI. Retailers are now playing catch-up.
One survey asked 1,348 US shoppers and 100 retail decision makers what this Christmas looks like, and the two groups described different seasons.
The gap, in two numbers
Narvar found 65% of consumers plan to use AI for at least one part of their holiday shopping. Only 8% of retailers describe themselves as very confident using AI to improve the shopping experience. That is demand already running in production against a capability still sitting in pilot.
Where the AI actually shows up
The uses are specific, not vague. Some 43% will use AI for gift discovery, 33% for comparing products and summarising reviews, 29% for budgeting the season and 15% for tracking and returns after checkout. Nearly eight in ten would use AI tools if it made shopping feel more personalised.
Retailers are watching a different screen
Only 14% expect AI shopping assistants to be the season’s biggest behavioural change, with attention still on shipping costs, discounts and operations. That blind spot costs money, because 56% of shoppers have delayed or avoided a purchase over refund uncertainty. Chief executive Anisa Kumar says the winners will meet what AI has already trained shoppers to expect.
Our takeaway: Put your AI shopping assistant into the Christmas plan now, because your customers have already put it into theirs.
📖 Read the full article at Chain Store Age
