8 Meaty Agentic Commerce News Stories from the Last Week that Revealed New Trends and Data That Flew Under the Radar
Fresh data on Amazon, ChatGPT, Gemini, Claude, AI Ad systems, Europe Agentic forecasts, Brands are falling behind and Can you guess P&G's Top Prio?
It was a chaotic week with all the Mythos and the SpaceX IPO news dominating all the headlines. In addition to the big Walmart Sparky and ShopAgentic news, we have curated the Top 8 Agentic Commerce stories that both delivered new strategic data-points and also flew under the radar this week.
Amazon Reveals Interesting Agentic Ad Details and new Datapoints.
Our new VP of Research, Andrew Bell noticed in this Amazon News article about Agentic Ads, they quietly dropped 2 new stats we’ve never seen before:
20% of shoppers who interact with a prompt continue the conversation about that brand
Adding prompts to a ‘sponsored brands’ ad drives a 6% increase in conversions.
The conclusion is that even though they are very new, we’re starting to see very promising ROAS data from Sponsored Prompts. This dovetails nicely with this one as we move to OpenAI’s ad system where this week we learned even more new stats→
Pacvue x OpenAI: 20% of ChatGPT Conversations Have Shopping Intent
Pacvue ran a webinar with OpenAI on ads, and it surfaced some first-party stats we hadn’t seen before:
20% of all ChatGPT conversations carry shopping intent
OpenAI frames the journey as Discover → Eval → Buy, with a “single product evaluation” step baked right into the funnel
OpenAI’s own comparisons show how much better ChatGPT performs vs. traditional search
My open question → if customers are discovering, evaluating AND buying inside ChatGPT, then why ditch the checkout? Something I question every day.
You can watch the webinar in it’s entirety here without a login→
Similarweb: Google/Gemini Keeps Chewing Away at ChatGPT’s Lead (and 2 more datapoints!)
Similarweb just dropped their May GenAI traffic report, and the trend line we’ve been tracking held → Google’s AI surfaces keep eating into ChatGPT’s lead, month over month. As a reminder, this is observed traffic data, not a survey — it’s what people are actually doing, not what they say they’ll do.
In their X post, Similarweb framed the trend this way:
12 months ago:
ChatGPT: 76.4%
Gemini: 8.9%
DeepSeek: 5.3%
Grok: 2.8%
Copilot: 1.9%
Perplexity: 1.8%
Claude: 1.6%
1 month ago:
ChatGPT: 52.7%
Gemini: 27.3%
Claude: 8.9%
DeepSeek: 4.0
Grok: 2.8%
Copilot: 2.0
Perplexity: 1.3%
Google started from behind, but this shows the power of distribution. Google is leveraging search, gmail, maps and every consumer touchpoint they have to not only stay relevant, but to definitely put a damper on ChatGPT’s aspirations of owning this market.
Also don’t sleep on number 3 - Claude is quietly growing with consumers at 9% I see increasing evidence of this every day. On weekends I wear a Claude hat and have been for 6mos - At first, nobody knew what Claude was now a large % of people recognize it and want to chat Claude.
Similarweb dropped two more nuggets in our world this week→
ChatGPT Advertisers grew 46% w/w! (2,585→3,770)!
Finally, probably the biggest datapoints that slide right under the radar that impacts Retailers the most: Google’s
On the surface this is pretty scary, we went from Google delivering 55% of their traffic in, right back ‘out’. Today that’s 32% and dropping daily. But it gets worse, when you look at the 32%→
Of that 32%, 27.4% of the traffic goes right back into Google -kind of an ‘in house ad’. 6% goes to paid ads (SEM) and 66% goes ‘out’ unpaid (organic/seo).
If you’re a retailer or brand with SEO+SEM, that’s a 72.6% shot on 32% of the traffic or a paltry 23.2% of Google traffic now coming ‘out’ that you have to share with everyone else in the mix.
Forbes + S&P Global: 55% of Merchants Say AI Agents Will Be a Major New Channel
Forbes ran a piece built on an S&P Global survey of 500 merchants, and they take a payments angle that’s worth the read. The underlying report is super-paywalled, but here’s what’s revealed in the summary:
65% of merchants strongly agree they’re considering adding a new payment partner to support agentic commerce
55% believe AI agents will become a major new transaction channel — my favorite stat in the whole piece
Two-thirds believe agents could initiate at least 10% of their e-commerce transactions within three years
94% cite pace of innovation as an important criterion when picking a payments partner
Top required capabilities → data privacy / regulatory compliance and trusted identity & authentication for AI agents
Here’s the question I keep coming back to: if agentic commerce is a hallucination, why are 55% of merchants leaning in? Are they hallucinating too? Or do they see something in the data that isn’t a hallucination? Naysayers often use AR/VR as an example of past ‘industry-wide mass hallucinations’, but I can guarantee you there was never a survey tht said
DCG Benchmark: The Gap Is About Execution, Not Technology
DCG — an independent digital commerce benchmarking firm with clients like Diageo, Mars, Henkel and L’Oréal — shared findings from their AI-in-digital-commerce benchmark across 13 product categories.
~30% are building genuine AI capability; only ~10% have generated real actionability; under 5% are deriving measurable value
90% have AI governance in place, but only ~35% have named a clear internal AI owner
60% believe AI can lower costs; only 38% can prove it today
15% can measure any incremental sales benefit from AI so far
No manufacturer scored above 50% on any agentic-readiness question
The pattern we keep seeing → the gap between AI leaders and everyone else is about execution and ways-of-working, not budget or tech. Fix the data layer (product data!) first, then layer AI on top.
Link → https://salsify.ondemand.goldcast.io/on-demand/cd56a56e-8630-4d6b-9741-367b071c3815
Sopra Steria: Agentic Commerce is A €310B European Opportunity
Sopra Steria put out the first European study dedicated to agentic commerce, based on a survey of 8,400 consumers across eight countries. They peg the potential European market north of €310 billion within ten years — but the headline is that the race won’t be won on technology alone:
41% of Europeans don’t yet trust any single actor to provide a shopping agent. Banks emerge as the most trusted (27%)
55% have already heard of agentic commerce; awareness hits 76% in Norway vs. just 38% in France
45% would delegate electronics & tech purchases to an agent, but only 16% for healthcare and groceries
74% still hit friction shopping online today
My takeaway from this is we have a “take of two Europes” with the North leading and the south lagging. The data indicates that trust is the gating factor. Right now nobody owns it, which is a wide-open field for whoever gets their first.
P&G’s CIO Lays Out the AI Playbook (35 Petabytes and Counting)
At the Evercore Consumer & Retail Conference, P&G CIO Seth Cohen spent nearly an hour detailing what may be the most sophisticated AI deployment in all of CPG. TIKR’s writeup is a great recap:
A proprietary ~35-petabyte consumer behavioral database, fed by 2M+ touch points a year and thousands of IoT-connected homes
“Digital twins” of consumers that have cut some R&D timelines from 5+ years to under six months
AI-driven, in-housed media: digital and social now need 10–200x the content volume — at higher impact and lower cost than agencies
GEO deployed across every category — they found the #1 source for Pampers info wasn’t Pampers.com, it was a Forbes forum thread
Supply Chain 3.0 running “unattended” portions of manufacturing shifts
What we have here is the CIO for the largest CPG telling us that one of their top priorites is GEO→
Visa: Intelligent Commerce, Agent Scoring & an OpenAI Tie-Up
At the Visa Payments Forum 2026, Visa rolled out a stack of agentic-commerce infrastructure under “Visa Intelligent Commerce.” The payments players are clearly racing to build the rails:
Agent Score → lets merchants test whether AI agents can navigate and complete tasks on their site (agentic-readiness, basically)
Agentic Directory → a registry of agents and merchants Visa has verified as legitimate
OpenAI partnership → secure Visa payments inside OpenAI’s agentic experiences
Large Transaction Model → an AI model trained on billions of transactions to cut fraud and reduce false declines
On the back end → stablecoin settlement now at a ~$7B annualized run rate, plus 160+ stablecoin-linked card programs live or in development
The top pull-quote is from the Visa Press Release and has caused me a lot of confusion.
Link → https://usa.visa.com/about-visa/newsroom/press-releases.releaseId.22491.html
Summary
That’s a lot to process, if we super up-level it, my take-away is across many dimensions - we’re seeing Agentic Commerce continue to grow and mature faster than anything we’ve seen from both the consumer and merchant side.
Retailgentic Editorial note: This week there are some news items coming and we don’t want to flood your inboxes, so there won’t be a Tuesday long-form, we’ll be back June 23 with Part II of 3-part series.
Great read.
ChatGPT = dilution, not decline.
Gemini is capturing this growth faster than it's adding users - queries now resolve inside Google's own surfaces. What’s exciting - brand visibility is a multi-engine discipline that can lead both the LLM horse and the shopper water.
We are about to spell “recommend” t-r-a-n-s-a-c-t