#BlinkTank: Google confirms paid AI testing, and a scam warning
It has felt like things have been moving at a breakneck pace lately, so this week actually seemed like a bit of a chance to catch our breath. It is nice to have a temporary reprieve from the constant "paradigm shifts" and "existential threats."
That said, there have still been some pretty significant stories landing. Most of the news this week - and many of the updates we are seeing across the platforms - aren't actually revealing anything new. Instead, they are largely confirming things we already knew were coming.
Google confirms paid AI testing
Alphabet’s latest earnings report showed Google Search hitting $63 billion in revenue, but the real point of interest for merchants was the official word on AI Mode.
Chief Business Officer Philipp Schindler confirmed that Google is in the "early stages" of experimenting with AI Mode monetisation. They are currently testing ads placed directly below AI responses. This surprises absolutely no one, but it confirms that the transition to an "Answer Engine" will be funded by the same ad-heavy model that has sustained Google for two decades.
A quick thought: if you’re worried about AI search destroying the ad model, don't be. Google will find a way to put a "Sponsored" tag on a hallucination if it helps the quarterly earnings.
📎 Search Engine Journal: Google details AI mode ad tests
The Shop App's parallel evolution
Harley Finkelstein announced a major upgrade to the Shop App this week, featuring a complete interface redesign.
We have always viewed the Shop App as an interesting parallel to the rise of AI search. It acts as a giant aggregator, pulling products from thousands of brands into a single, unified feed. This update confirms that Shopify is doubling down on this aggregated discovery model. It is training users to browse products in the exact way that LLMs are now starting to present them - moving further away from the traditional "list of links."
📎 Harley Finkelstein: Shop App's huge upgrade
The ChatGPT merchant dashboard is in test mode
Thanks to Juozas Kaziukėnas for sharing a preview of the new ChatGPT eCommerce Merchant dashboard. While it is currently limited to a small group of trial users, it confirms our suspicions about how OpenAI intends to handle commerce.
The backend is built entirely around a product feed. While there is no performance reporting yet, we expect that to follow quickly. Shopify is already working on its own AI search dashboard, and Google’s Merchant Centre is moving in the same direction.
📎 Juozas Kaziukėnas: A preview of the ChatGPT merchant dashboard
A warning on OpenAI phishing attempts
Last week, we mentioned that OpenAI is asking for a minimum $200,000 commitment for its beta ads program. As predicted, scammers have moved in to exploit the high entry barrier.
Marketers are being targeted with phishing attempts via Apple’s TestFlight, offering "beta access" to a fake OpenAI ad platform. These are fraudulent attempts to gain access to accounts or data.
📎 Scammers exploit OpenAI’s name in fake ad platform targeting
LLM-specific files: A waste of effort?
There has been significant chatter about creating specific markdown files or .well-known directories to help LLMs "understand" your site.
Lily Ray recently shared official responses from both OpenAI and Google that confirm our internal view: creating bespoke markdown files for LLMs is largely a waste of time. These models are built to understand the web as it is - via HTML, schema, and structured data. Adding separate files specifically for AI likely adds more work for little benefit.
📎 Lily Ray: Official responses on markdown files for LLMs
A rare update to Google Discover
This is a slightly unusual one: Google has released a specific documentation update regarding the "Discover Core Update".
While Discover is traditionally a driver for informational and publisher content rather than direct eCommerce, any update to the algorithm is worth noting. It is a reminder that Google is constantly recalibrating how it pushes content to users before they even search. For brands with a strong content strategy, this update could shift the baseline for top-of-funnel organic traffic.
📎 Google Search Central: Discover core update documentation
From the blog: Taxonomy as the foundation
If structured data is how LLMs understand your site, then taxonomy is the language they are reading. Taxonomy is foundational to everything we do at Blink, yet it remains one of the most misunderstood parts of e-commerce.
We have put together a guide explaining what a taxonomy actually is, why it is so critical for Shopify stores, and how a well-structured hierarchy unlocks growth across every channel - from cleaner SEO to more efficient PPC feeds.
📎 Why taxonomy is so important for Shopify stores
Network shout out: PyRetailScience
This week we want to give a shout out to Murray Vanwyk and his work on PyRetailScience.
Murray is a hugely experienced retail data scientist who has distilled complex methodologies- the kind used by massive consultancy groups like PwC - into a Python package for retail data science.
This is the kind of high-level data science work that major high-street retailers use for analysis, and Murray has made it incredibly accessible. It is a very cool project and something more people in the space should be taking notice of.
📎 Visit PyRetailScience
And that's it for this week!
As ever, thank you so much for reading. We've had some wonderful feedback about this newsletter over the past few months, and we genuinely appreciate it.
All the best,
Sam