Cloudflare just gave ai visibility tools a free scoreboard, and nobody can prove what moves it yet

Cloudflare turns ai visibility tools into a dashboard everyone can open#

A fox does not wait for the hare to wander into the open under a moonlit hedgerow; it reads the hush for the faintest disturbance and moves on a patient prowl before the scent goes cold. That instinct is roughly what sits behind Cloudflare's newest release, because on 6 August 2026 the company folded its existing Agent Readiness checker into a broader AEO Suite and switched on an early-access AEO Visibility Dashboard, the freshest name in a fast-growing field of ai visibility tools now sitting inside a CDN's own control panel rather than a separate subscription. Cloudflare's own announcement frames the shift bluntly: the moment a customer chooses you now often happens inside a model's answer, before a human ever sees your homepage.

The number behind that claim is worth sitting with. By Cloudflare's own count, "fewer than half of all HTML page requests now come from a human." Not every one of those requests is an agent acting on a shopper's behalf, but independent reporting backs the direction of travel: Cloudflare's own traffic split crossed over in early June 2026, with 57.4% of requests to Cloudflare-hosted sites arriving from bots against 42.6% from humans, a milestone chief executive Matthew Prince had pencilled in for 2027. Prince's own explanation of why the curve bent so sharply is worth keeping: a human might browse five sites before buying a camera, while an agentic assistant, per his SXSW 2026 remarks, might browse five thousand.

That gap is the thicket every marketing team is now hunting through, and it explains why interest in ai visibility tools, and in any dedicated ai visibility tracking tool sold alongside them, has climbed sharply within eighteen months. Cloudflare's pitch is that visibility should not require a separate vendor at all: fold the score into the same dashboard already serving the traffic. It is a genuinely useful instinct, and on the evidence below, it is also an early one.

Two products now sit under one roof, and folkfox thinks it is worth being precise about which is which before they blur into a single line item on somebody's invoice. Agent Readiness asks whether a machine can reach you at all: permission to crawl, a sitemap to navigate, content it can actually parse. The AEO Visibility Dashboard asks the harder question further down the trail: given that a machine can reach you, does it recommend you. One is plumbing. The other is reputation. Confusing the two is the single easiest mistake a client will make with any of the newer ai visibility tools this quarter.

Inside Agent Readiness: the ai visibility tools foundation Cloudflare folded in#

Cloudflare's diagnostic half of the AEO Suite is not new; it is a rebadge of Agent Readiness, which the company shipped on 17 April 2026 as a public scanner at isitagentready.com, and it deserves its own scrutiny before anyone touches the citation numbers, because a site that fails the plumbing test will never generate an honest reading on any ai visibility tools that measure reputation. The scanner groups its checks by effort: quick wins such as a crawler-readable robots.txt, an XML sitemap and AI-crawler rules; technical groundwork such as Content Signals and an API catalogue; and advanced integration such as OAuth discovery, MCP server cards and Markdown for Agents, which lets a crawler send an Accept header for text/markdown and receive a converted document instead of a page stuffed with navigation and script.

How agent-ready the web actually is#

The scale of the gap is the real story, and it is a genuinely useful reading for anyone shopping among today's ai visibility tools on the promise of easy wins. Cloudflare's own Radar data, published alongside the April launch, found robots.txt nearly universal at 78% of sites, though the company is careful to note that most of those files were written for search crawlers, not AI agents. Content Signals, the newer standard for declaring how content may be used, appears on only 4% of sites. Markdown content negotiation passes on 3.9%. MCP server cards and API catalogues together, per the same Radar count, sit on fewer than 15 sites in the entire dataset.

Cloudflare's own developer documentation practises what it scores: its pages carry a short banner telling any crawling model to fetch the Markdown edition instead, on the reasoning that HTML "wastes context," and the Agent Readiness API now scans domains weekly so the same figures can be pulled programmatically rather than read off a chart. That is a genuinely useful bit of scaffolding, and it is also a reminder that readiness here means machine-parseable, not persuasive. A page an agent can read cleanly still has to earn a citation once the model is forming an answer, which is exactly where the newer half of the AEO Suite picks up the trail.

Independent commentary on the scanner is a useful corrective, and it cuts in the direction folkfox would expect from a first-generation entry among ai visibility tools. Search Engine Journal tested the composite score against itself and found the same site returning 33 out of 100 under an "All Checks" preset and 67 out of 100 under a "Content Site" preset, a swing wide enough to make the headline number close to meaningless without reading the checks beneath it. The piece's warning is blunt: "when a measure becomes a target, it stops being a good measure."

Citation Rate, Mention Rate, Prominence, Share of Voice: the new ai visibility tools vocabulary#

The genuinely new half of the release is the AEO Visibility Dashboard, and its four terms are about to become the shared vocabulary of every one of the ai visibility tools now pitching a marketing budget. Citation Rate is the share of category answers that cite a site as a source. Prominence covers, when a site is cited, how much of the answer is actually theirs and how early it lands. Mention Rate tracks how often an assistant names a brand at all, whether or not the site is the cited source, which separates simple awareness from earned attribution. Share of Voice is a brand's slice of citations against competitors, showing who is winning the prompts a rival is losing.

Cloudflare is explicit about method, which is the detail folkfox would ask any vendor to disclose first: "we probe the leading assistants (today, Anthropic's Claude and OpenAI's GPT)" with prompts built to mimic realistic customer discovery inside a given industry and category. That is two assistants, not the whole answer-engine landscape, and it is worth saying plainly: a dashboard that never queries Perplexity, Gemini or Bing Copilot cannot claim to cover the whole quarry, only two of its busiest tracks.

A crowded thicket of paid rivals#

It also is not launching into an empty field. A crowded market of ai visibility tools has raised real money on the promise of the same four metrics, useful context for any client wondering whether a free CDN dashboard threatens their existing subscription or simply offers a second opinion worth having. Independent comparison reporting from May 2026 put Profound at $155m raised and a $1bn valuation after a February 2026 Series C, Peec AI at $29m raised with more than $4m in annual recurring revenue inside ten months, and Otterly's entry tier at $29 a month.

Set a free, early-access dashboard against that field of paid ai visibility tools and the disruption is obvious, even before the accuracy question is settled. Clients will not wait for the measurement science to catch up before screenshotting the number for a board slide, so the professional response is not to dismiss the reading but to know precisely what it can and cannot support.

Why no ai visibility tools reading is proof of anything, yet#

Here is the honest, slightly uncomfortable centre of this piece, and it is where folkfox's own evidence file earns its keep. The most rigorous academic look at this whole category so far is unambiguous, and it applies just as much to a free CDN dashboard as to any paid rival among today's ai visibility tools. A critical survey covering three years of generative engine research states plainly that "no reviewed technique shows a stable, longitudinal, cross-platform causal effect on organic discoverability." The same survey goes further, in a line that applies directly to the very kind of tool Cloudflare just shipped: "Commercial audits further reveal low source overlap, substantial run-to-run variability, and persistent fidelity gaps."

Read that twice against the excitement of a new free score. Direct network visibility into who is crawling a site does not exempt any dashboard from measuring an underlying system that Google itself documents as variable between runs, nor from the instability researchers keep finding when they repeat the same prompt more than once. A 2026 study that ran 401 prompts across 4,812 responses found repeated identical prompts returned inconsistent answers in up to 21% of cases, per Encarnación and colleagues. A single AEO Visibility sweep, however slick the dashboard around it, is one draw from a moonlit measurement, not a settled reading of where a brand stands.

None of which makes the category worthless, and this is where the useful signal actually lives, buried under the noise everyone else is chasing. The largest controlled citation study to date, 252,000 trials across six large language models and eighteen content factors, found "topical relevance and list position are the biggest drivers of being cited first", while formatting-only edits barely moved the outcome at all. A separate framework analysing 602 controlled prompts and 21,143 citations across ChatGPT, Google's AI features and Perplexity found something sharper still: being cited and being absorbed into an answer are different prizes, and the pages that actually shape a response are "longer, more structured, semantically aligned, and richer in extractable evidence", dense with definitions, numbers, comparisons and named procedure.

A short, honest aside belongs here too, because two Cloudflare products landed within days of each other this month and they are not the same story. On 4 August, two days before the AEO Suite update, Cloudflare also switched on stablecoin wallets that let AI agents pay for goods and services autonomously, complete with spending caps and human-readable handles. That is agentic-commerce infrastructure, a payment rail, not a citation-tracking product, and folkfox would gently steer any client deck away from blurring the two into one "Cloudflare is doing AI stuff" slide. They share a vendor and a launch week. They do not share a metric.

Turning a free ai visibility tools reading into something defensible#

So what does a defensible practice with ai visibility tools actually look like this week, while the measurement science is still catching up with the marketing. Treat the AEO Visibility Dashboard, and any of the other ai visibility tools sitting alongside it, the way a tracker treats fresh footprint in soft ground: real evidence worth following, not proof of what made it. Run the free reading, because it costs nothing and it is genuinely informative directionally. Do not carry a single sweep into a board meeting dressed as a settled trend.

Pair the plumbing work with the content work, because the evidence above is consistent on where the real leverage sits. Fixing Agent Readiness failures, a missing sitemap, absent Content Signals, no Markdown negotiation, removes the reasons a model cannot read a site. It does nothing to make the model prefer that site once it can. That second job belongs to content marketing: specific, checkable, well-positioned passages, the kind our own SEO and GEO practice builds around relevance and position rather than formatting tricks, because Google's own guidance points at the same non-commodity content the controlled research keeps confirming.

For regulated clients the stakes sit sharper still. A fintech brand or an iGaming operator cannot afford a model mis-stating a licence, a rate or a rule inside a confident-sounding answer, so treat Prominence and Mention Rate as compliance signals worth watching as closely as any growth metric. Our own reporting on why a single reading is a rumour, not a measurement sets out the cadence discipline that makes any of these ai visibility tools worth paying for in the first place, and it sits alongside what we found when a model cites a brand everywhere and recommends it nowhere, and the pattern behind why models search for the brands they already know.

The fox that outfoxes the hunt is not the one that finds a single fresh scent and calls it proof. It is the one that keeps checking the wind and wariness of the hedgerow, night after night, until the pattern holds and the trail back to the den means something. Cloudflare has just handed every marketing team a free nose for that hunt. What folkfox is here for is teaching a client what any one scent from today's ai visibility tools is actually worth this week, and what it will take before it is worth repeating in a board deck without a footnote.

Frequently asked questions#