At What Cost?
You probably don’t need to be told that grocery prices have skyrocketed since the pandemic – up 25% in five years. But the bad news only continues: Retailers are seeing how far they can push the envelope on dynamic pricing.
Instacart has been experimenting with selling the same product from the same store at different prices, which a nonprofit determined by enlisting volunteers to add the same items to their cart. Eggs, for example, varied by 20%, from $3.99 to $4.79.
But why? An Instacart spokesperson said it changes prices to “[help] retail partners understand consumer preferences and identify categories where they should invest in lower prices,” a spokesperson tells The New York Times. But one retailer, Target, said that only Instacart determined prices on its platform. Others declined to comment.
The rise of ecommerce and evolving technologies have made pricing “more flexible,” says Harvard economist Alberto Cavallo, allowing for a rise in dynamic and surge pricing (like Uber’s model) and personally tailored pricing.
Different prices for different people isn’t a new strategy, points out Yale economist Kevin Williams. Discounts for students or senior citizens and all sorts of loyalty programs have existed for generations. But those strategies are transparent to all involved.
Instacart’s personalized pricing, unlike personalized advertising, does not take into account a user’s behavior or personal attributes, the company claimed. When companies appear to cross that boundary – as Delta airlines did this summer when it boasted of AI pricing – consumer and regulatory backlash has been swift.
Taking A Stand(ard)
One of the biggest challenges with a new technology boom is getting competitors on the same page.
Several leading AI companies, including Anthropic, OpenAI, Google and Microsoft, are coming together to develop a set of technical open-source standards to be implemented across agentic AI tools.
(Making agentic tech work for programmatic is another kettle of fish entirely.)
The group, called the Agentic Artificial Intelligence Foundation, is initially focusing on three existing open-source AI tools, reports The Information.
One existing open-source standard is the model context protocol (MCP), which is how AI agents connect to other applications. The second is Agents.md, which determines how best to give instructions to coding agents, and the third is Goose, an AI agent that runs locally and doesn’t require an internet connection.
Even among popular open-source tools, like MCP, the technology still has a ways to go. Security is a main concern for companies developing these tools, and they’re working to ensure that agents won’t be able to connect to compromised apps.
And, who knows, maybe one day Agentic Artificial Intelligence Foundation will standardize ad delivery across agentic interfaces.
Roll With It
Here’s one new training data set that could be a huge boon to Tinder: your photo roll.
The app has begun generating Photo Insights, as it calls the product, which is an opt-in option that scans a user’s photo roll to identify interests or preferences that could be helpful in connecting potential dates. [H/t Jane Manchun Wong.]
Photo Insights could even identify political or religious beliefs, Tinder has disclosed.
And what’s more, Tinder may show certain photos to a potential date. Someone with an expressed interest in skiing, say, might be shown photos of someone they’ve matched with on the slopes to underscore the lifestyle connection.
It’s a small news item and not exactly relevant for programmatic traders. But it shows the kind of unanticipated new sources of reference data that are uncorked by the introduction of an AI tool. And why brands love apps with reliable access to people’s camera rolls.
Also, there will surely be some humorous foul-ups involving photos shown to potential matches because of some connection made by Tinder’s AI systems.
But Wait! There’s More
McDonald’s pulls its AI-generated Christmas ad for the Netherlands down from YouTube after intense backlash. [Futurism]
Instagram is generating inaccurate SEO headlines for posts that appear in Google searches. [404 Media]
Telemundo has already sold 90% of the ad inventory for the FIFA World Cup in 2026. [Variety]
Advice for publishers considering adding paywalls. [Search Engine Journal]
Why the most effective holiday ads aren’t new at all. [Cynopsis]
Published book authors are increasingly being targeted by AI-powered marketing scams. [Electric Literature]
Gemini is catching up to ChatGPT in the battle for AI search dominance. [The Information]
You’re Hired!
Keynes Digital appoints its first CTO in Sachin Telang. [release]
Thanks for reading AdExchanger’s Daily News Roundup. Want it by email? Sign up here.
