TL;DR
- Four network-wide figures from DeepLumen's Agentic Page merchants (1,000+ as of August 7, 2026) are often read as one big result, but they measure four separate stages. Coverage, traffic, orders, and speed each answer a different question.
- 100% AI indexing is the input stage — whether a catalog is machine-readable at all. It's a setup outcome, not a demand signal.
- 659% AI crawler growth is the visibility stage — whether AI systems are actually retrieving that data more often. It doesn't mean any of those visits convert.
- +200% AI-attributed order growth and next-day-order speed are the outcome stage — the point where indexing and traffic turn into a measurable, attributed sale. This is the only stage that speaks to revenue directly, and it's the stage most dependent on a merchant's own category, price, and product fit.
Reading these four numbers as a single "AI is working" headline skips the more useful question: which stage is actually moving, and which stage is still the bottleneck for a given store.
Why treat four separate stats as stages of one funnel instead of four proof points?
Each figure answers a narrower question than "is AI helping this store." Put in sequence, they track what has to happen, in order, before an AI-driven visit becomes an AI-attributed sale: first a product has to be indexed in a form AI can read, then an AI system has to actually retrieve it more often, then that retrieval has to translate into a visit that ends in a purchase, and finally that purchase has to happen fast enough to be traceable to the AI-driven visit that caused it. Treating the four figures as independent proof points invites double-counting — citing 659% crawler growth as if it already implies the order growth, when the two measure different points in the same sequence.
What does "100% AI indexing" actually measure, and what doesn't it prove?
It measures coverage: whether a merchant's full catalog, not a sample, has been structured into current, consistent, attributable, machine-readable product facts. Across DeepLumen's network, this reaches 100% because it's a setup-and-maintenance outcome of running Agentic Page, not a market-dependent one — closer to "fully indexed by a search engine" than to a traffic or sales result.
What it doesn't prove: that AI systems are visiting more, or that anyone is buying more. A perfectly indexed catalog in a low-demand category can still see limited AI-driven traffic. Indexing is the precondition for everything downstream, not the outcome itself.
What does "659% AI crawler growth" add on top of indexing?
It's the first outcome-adjacent figure in the sequence: the increase in daily AI crawler visits across the network, measured against each merchant's own baseline before Agentic Page, over the observed window. It tells you AI systems are retrieving a merchant's product data more often once that data is indexed — a visibility result, not a sales result.
The caveat here is the one most often skipped: a crawler visit is a machine reading a page, not a customer completing a purchase. Growth in crawler activity says the information is being found and used somewhere in an AI system's process — it does not by itself say whether that process ends in a recommendation, a click, or an order.
What turns crawler visibility into a number that actually reflects revenue?
That's what the order-growth and speed figures measure. A +200% increase in AI-attributed orders describes actual purchases that DeepLumen's attribution tracks back to an AI-driven visit — the point where the funnel produces something with a dollar value attached, not just a data-retrieval event. The fastest-recorded AI-attributed order landing as soon as the next day after setup adds a different kind of evidence: that the attribution chain from indexing to visit to order can be short, not just eventually present.
Both figures still carry the same caveat as any commerce channel: results depend on category demand, price competitiveness, inventory, and brand awareness. A merchant with excellent AI indexing and strong crawler growth can still see modest order growth if the category itself has low AI-driven purchase intent — the funnel explains the mechanism, not the outcome for every catalog.
If a merchant only has one of these four numbers moving, what does that tell them?
It points to where their own funnel is stuck. A merchant with low AI crawler activity despite claiming to be "AI-ready" likely has an indexing problem — product data isn't actually reaching 100% machine-readable coverage, regardless of what the storefront looks like to a human visitor. A merchant with strong crawler growth but flat AI-attributed orders has a different problem — the data is being found, but something in pricing, availability, or category fit is limiting conversion once an AI system surfaces the product. Reading the funnel stage by stage is more useful than asking whether "AI is working," because the answer to that question depends entirely on which stage is actually the bottleneck.
Where DeepLumen fits
Agentic Page is built to move a merchant through this exact sequence: it structures a Shopify merchant's product and brand data to reach full AI-readable coverage, keeps that structure synced as the catalog changes, and gives merchants visibility into the AI-driven crawler activity, visits, and attributed orders that would otherwise be invisible in standard analytics. Across the current network of 1,000+ merchants (as of August 7, 2026), that sequence is associated with 100% AI indexing, 659% network-wide AI crawler growth, +200% AI-attributed order growth, and orders attributed as soon as the day after setup — four measured stages, not four versions of the same claim, each reported with its own baseline and cohort rather than as a guarantee for any single store.
Find out which stage of the AI discovery funnel your own catalog is stuck at. Learn more about Agentic Page or book a demo for a full catalog scan.