🚨DATA DROP🚨: The ReFiBuy Q2 AI1000 Is Out! Self-Service Portal, Big Four Highlights and Are You Ready for Agentic Holiday 2026??

We've got category data, UCP details, answer engine %-ages and much much more, let's get into it...

In early 2026, I had what I’d call ‘a wacky founder’ idea. The kernel of the idea goes back to my ChannelAdvisor days, specifically 2004, when Internet Retailer Magazine (!!) launched the first IR300 – a ranked listing of the top 300 retailers selling online with this new-ish thing called: “e-commerce” (obviously a fad!!).

Both retailers and vendors were excited. Retailers loved seeing the benchmarking data and what vendors everyone was using. As vendors it was super helpful showing the world that we were so dominant in the ‘marketplace management category’. From there the list expanded in depth (ultimately stopping at the IR2000) and width (different countries, regions, etc.). IR was acquired, the remaining piece separated from the tradeshow, was re-branded Digital Commerce 360 but the IR1000 list is still an annual publication and on its 22nd year of publication (available here).

Back to the wacky founder idea… That 2004-inspired idea was: “What if we took the IR1000 and we added multi-dimensional benchmarking data for Agentic Commerce?”. The DC360 folks loved the idea and we collaborated for ~6mos and in July, 8 weeks after the 2026 IR1000 dropped in May 2026, we released the Q1 AI1000.

But…my pitch had one important caveat that came along, a ‘must have’ if you will: Because the underlying AI models and Agentic Commerce is evolving so quickly, we had to update the AI1000 quarterly. In 2026, annual updates feels like the difference between X and Y (5yrs). This was a big investment on our side (people, tokens, code, infrastructure) , we had to not only build the infrastructure for the Q1 AI1000, but going forward, our aspiration was that days after the end of a Q, we’d be in a position to drop the next quarterly update.

The Results?

The response to the Q1 AI1000 exceeded our wildest expectations. We’ve had a ton of positive feedback, and a lot of retailers asking all the important questions. Most importantly we’ve been flooded with awesome feature requests:

• Can I update this in more real-time – we just implemented UCP and want it to show up!

• When Q2 comes out, can I be alerted? Can I login and look at the six-factors and trends at that level for our company?

• We have four sub-brands, can we track those separately?

• Can I create a list of companies I want to track on a watch list?

Unfortunately, Q2 took a bit longer than we wanted. There’s two main reasons.

1. We had a large influx of customers at the end of Q2 (which we’re hugely thankful for 🙏) and had to prioritize customer on-boarding vs. AI1000 work.

2. Being Agentic, once we got onto the Q2 AI1000, the team wanted to make it worth the wait and pushed themselves to add all (and more) of the features above (minus the ‘watch list’ we’re looking at that for Q3).

I’m particularly excited by the self-service portal so let’s start there.

Introducing the AI1000 Self-Service Retailer Portal

The way this works is you have to be a team member of the company. For example, let’s say you are in marketing at Chewy. You go to refibuy.ai/ai1000/chewy (where chewy can be any of the AI1000 company names) and you enter your email. We enter a magic login link to that email to verify you are with the company. Click the link and you are off to the races. To give you an idea of what it looks like here’s the top part of Chewy’s data:

Note formatting modified for this newsletter format, some details redacted

Not only do you get complete insight into the four factors that combine to form the score that determines your AI1000 rank, but with the Q1->Q2 slider (which in Q3 will have a third datapoint and so-on), you can track all your detailed progress Q/Q and ultimately Y/Y, etc.

Also - you’ll see the 90-day momentum, that’s coming in Q3 when we have a large enough corpus of trend data.

You asked, we delivered! I hope everyone enjoys seeing their rankings, if you aren’t on the AI1000, let us know, we can run a similar report for you, free of charge.

Now for some of my personal highlights from the AI1000 Q2 data drop, many more can be found that the team put together at our comprehensive landing page here and today’s press release here.

Four AI1000 Q2 Highlights

The first highlight I wanted to share illustrates why quarterly updates are critical in the AI Era: 972 of the 1,000 retailers changed positions from Q1-Q2!

This graphic illustrates the ‘rank move cohorts’ from downward moves to upward:

In the 119 ‘top climbers’ club, here’s the 8 companies that moved the most:

Congrats to these AI1000 top movers, your hard work is paying off!

Category Trends

Peeling the onion on all this individual retailer movement, here’s an illustration of how each category moved:

Congrats to all our friends in the automotive category -you guys are really pushing hard and it’s really paying off.

That’s the Q1->Q2 story, here’s where the categories rank stack as of Q2:

Answer Engine Moves

You’ve probably all seen the Similarweb chart that shows the dramatic loss of share by ChatGPT of referral traffic (but overall traffic for all players is still up substantially). In that data you see Gemini surging, but also many have noticed that Perplexity is fading and Claude is on the rise – yes Claude, with it’s enterprise audience, consumers are flocking to Claude:

Claude hasn’t surpassed Perplexity in our dataset, but one more Q like this and that will happen.

UCP Adoption

One of the big factors we measure is what retailers have a UCP endpoint. Sadly, only 225 of the AI1000 (22.5%) are there. The bulk of these picked up the UCP endpoint ‘for free’.

Want to move your score for Q3 – Implementing UCP is your easiest and fastest unlock.

The Clock Is Ticking Toward Q4

We’re down to 14 weeks until Thanksgiving, there are four things you can do in those 3 months to be ready for what we believe is going to be a huge Agentic Holiday. We’re not alone, Salesforce just dropped their Holiday Predictions (pod coming soon!) and their second prediction is that AI agents will influence 22% of sales this holiday or $73B in Agentic Commerce fueled opportunity for you to grab.

1. Always Updated Product Data - Enrich your product data for agents and keep it continuously ready for agents to answer questions agents will be asking.

2. UCP - Implementing agentic commerce protocols give you an unfair advantage

3. Optimize for a multi answer-engine market – The big two are ChatGPT and Gemini, and you can get them with datafeeds and ACP/UCP endpoints. From there, the other engines are crawler based, so take that awesome product data you’re exposing and put it in your PDP-level meta data.

4. Retail is a Zero Sum Game - If your score (available in the portal) is below your category average, you’re not lagging your competitive set, time to really focus on getting caught up, because at this pace of change, if you fall too far behind, it will get exponentially harder to catchup.

Summary

The full Q2 findings, category analysis, and the new Retailer Portal are live at: https://www.refibuy.ai/ai1000 .

That’s just the tip of the iceberg, keep an eye on our ReFiBuy social channels, we’ll be sharing even more insights and analysis over the coming weeks. And of course, in less than 90 days we’ll be back with the Q3 data, the last AI1000 ‘print’ before Holiday 26. Then in early 2027, we’ll close out the year with a recap of how everyone did with Agentic Holiday 26 with the Q4 data.

I want to wrap by thanking Digital Commerce 360 to listening to my wacky idea (and then saying: “Let’s Go!!”), the AI1000 retailers who gave us so much feedback from the Q1 data, our media partners and friends who help spread the word, our amazing customers for their support and the team at ReFiBuy for working tirelessly to get this out and deliver an amazing set of new features on top of the data – enjoy everyone!