Agents that shop your store in a real browser
Not a guess from your HTML: AI agents walk from product page to checkout, and you can watch the recorded run step by step.
Watch a recorded run →
AGENT READINESS INDEX 2027Free data
Your next customer is an AI. Is your store ready for it?
Specoria measures how shopping agents like ChatGPT, Gemini, Claude and Perplexity read your catalog, policies and checkout, and gives you ready-made fixes, each verified by re-measuring.
No sign-up · no email · no card
See how it worksOr send AI agents to your store (simulator)
Not an online store? Hotel →Service business →Agency →
01 · ScanREAL SCAN · web-scraping.dev · October 8, 2026
Free testweb-scraping.dev
- PassedSearch and user agents allowed (robots.txt)
- PassedProduct name and price readable without JavaScript
- PassedProduct structured data (schema.org)
- PartialProduct identity (GTIN/MPN, brand, SKU)
- FailedMachine-readable return policy
- FailedMachine-readable shipping information
75/ 100Partly ready23 checks measured
27 checks on public pages, no sign-up
02 · SimulateREAL RUN · STORE NAME HIDDEN · October 8, 2026
FindProductSizeCartCheckout
- 1Searched the store; the page came back in German, prices in EURDE
- 16Switched the site to Turkish itselfDE
- 22No add-to-cart button: stopped on the product pageTR
22 steps · never reached the cart · See the whole run →
An agent shops in a real browser and stops at checkout
03 · VisibilityILLUSTRATIVE · AI VISIBILITY RUNS ON PAID PLANS
“Which store has a good linen shirt under $60?”
- ChatGPTNamed and cited
- PerplexityNot named
- GeminiNamed, not cited
Do AI answers name your store?
Example answers, not a measurement: 8 questions × 3 engines are measured monthly on paid plans.
04 · TaskFROM THE SAME SCAN · CALCULATED, NOT RE-MEASURED
TASK · OPEN
Machine-readable return policy
+5 points when it passes
Fix the top 3: 75 → 89
Calculated by the score model; run the free test again to measure it.
Each finding becomes a task with code and a re-check · Open the sample report →
Source: Specoria free test, sample report ·
PRODUCT GRAPH
•product
•offer
•policy
•checkout
GPT
Gm
Cl
Px
Sh
ik
Wc
Tc
PLATFORMS SPECORIA MEASURES TODAY
- Shopify
- ikas
- Ticimax
- IdeaSoft
- T-Soft
- WooCommerce
- Magento
- Google Merchant
HOW TO START
Three steps. The first takes seconds.
Free test
Instant, no email: 27 checks on your public pages, with the result right here.
Run the free test →Free deep audit + panel
Leave your email: a PDF made in a real browser, plus a panel with tasks and ready-made code, within minutes.
Open my free panel →Starter, first month $1
Starter keeps measuring: re-checks, monthly AI visibility and alerts. Cancel from the panel in one click.
See pricing →
How deep each step goes
- Free test27 checks, no email
- Deep auditan agent’s 8 steps, up to 12 product pages, a real browser
- Panelthe full 196-criteria picture
Side path: Agent simulator. Free: watch AI agents shop your store live in a real browser. Watch the agents →
NEW · MYSTERY SHOPPER AGENTS
Send the agents →
Watch AI shopping agents try to buy from your store
Two agents shop your store live in a real browser and stop at checkout. See every step, where they get stuck and how to fix it.
- Free, no email
- Real browser
- About 5 minutes
PRODUCT TOUR
Specoria in 15 seconds
Watch Specoria read your scattered store data, show the gaps in one product record, and the fixed store come out on top in a shopping agent’s decision.
- Scattered data. Product pages, structured data, policy pages and the feed disagree, and some of it is missing or stale.
- Reading. Specoria reads your public pages, robots.txt, sitemap and product feed; no access to your systems.
- Product graph. One product record with every field linked to its source; missing and conflicting fields become tasks with fix code.
- Agent decision. The shopping agent compares sellers, and the store with every criterion verified comes out on top.
00:07 / 00:15
Concept demo: the interface and figures are illustrative, not a real customer result.
THE PROBLEM
Your product data lives in ten places. No agent can read it all at once.
TASKS THAT NEVER REACHED CHECKOUT · OUR 40-RUN REAL-BROWSER STUDY
22%
Lighter AI models failed to reach checkout in 22% of shopping tasks (2 of 9); the strongest reached it 9 of 9. The gaps came where the store made the path harder, such as a hidden checkout button or a hard-to-find cart link. Read the field study →
WHY NOW, IN NUMBERS
Where shopping starts is changing
AI assistants are becoming the first stop for finding and comparing products. Most of the figures below measure the US market (McKinsey’s estimate is global); read them as a leading indicator.
- +138% Growth in AI-referred visits to US retail sites in May 2026 versus a year earlier Source: Adobe Analytics (Digital Commerce 360) · Jun 17, 2026 ↗
- 14x AI-referred retail traffic since October 2024 (up 1,324%), across more than 1 trillion visits analyzed Source: Adobe Analytics (Digital Commerce 360) · Jun 17, 2026 ↗
- +54% Higher conversion rate for AI-referred visits than for visits from other sources Source: Adobe Analytics (Digital Commerce 360) · Jun 17, 2026 ↗
- 51% Share of US consumers who used at least one AI tool to shop last month; the first time above 50% Source: NIQ Agentic Commerce Tracker (The Next Web) · Sep 24, 2026 ↗
The figures are published by the organizations named, not by Specoria. Forecasts vary by organization. All sourced statistics →
WHY SPECORIA
Four things you can check for yourself
A free test with no email
27 checks on your public pages. No sign-up, no card; the result opens on its own page.
Run the free test →We never count what we can’t measure as a pass
Our weights are published. A check we couldn’t verify is left out of the score instead of being counted as a pass or a zero.
Read the method →Fix guides for your platform
Ready-made code with steps for Shopify and WooCommerce, plus ikas, Ticimax, IdeaSoft and T-Soft. Every fix is re-measured before it counts.
Open the platform guides →Not a store? We also measure Hotel → and Service business → with checks of their own.
METHOD
A living map of your store, built for agents
Specoria measures how your store looks to shopping agents, turns every finding into a task with ready-made fix code and re-measures to confirm each fix.
01 · READING SOURCES
REPRESENTATIVE
- TXTrobots.txt · sitemapREAD
- LDStructured dataREAD
- WEBShopify store412 MISSING
- PDFReturn policyCONVERTING
- FEEDMerchant CenterSTALE · 19 H
PRODUCT · VERIFIED
Espresso Pro X2
- GTIN8690000123456
- MPNEPX2-EU
- PRICE€449
- STOCK42 · live
- SHIPPING2 days · DE
- RETURNS30 days
- WARRANTY2 years
- WORKS WITHPods + beans
Merged from 7 sources · 2 conflicts resolved · every field linked to its source
Find me an espresso machine under €500 that ships in two days and comes with 30-day returns.
SHOPPING AGENT · 7 sellers compared
- 1Espresso Pro X2 · your storeAll criteria verified€449
- 2Competitor ADelivery time unclear€429
- 3Competitor BReturn terms unreadable€439
source: product page JSON-LD · delivery (DE-BE) → 2 days
INTEGRATIONS
Integrations: what we support today
Store platformSupported today
Shopify
Measures your store through an agent’s eyes and gives fix code and steps for your Shopify theme and admin.
Store platformSupported today
ikas
Prepares shipping and return terms for the Turkish market as machine-readable code, with a read-only guide for ikas.
Store platformSupported today
Ticimax
Measures how readable your product pages are to agents and fills missing structured data with ready-made code.
Store platformSupported today
IdeaSoft
Detects the platform automatically, measures product, policy and checkout steps through an agent’s eyes and prioritises fixes for IdeaSoft.
Store platformSupported today
T-Soft
Detects the platform automatically, measures structured data, shipping and returns, and opens gaps as tasks.
Store platformSupported today
WooCommerce
Gives guest checkout, robots.txt and structured data steps for WooCommerce and your SEO plugin.
Store platformSupported today
Adobe Commerce (Magento)
Detects the platform automatically and measures product data, add-to-cart and the checkout path through an agent’s eyes.
HotelsSupported today
Hotel booking engines
Recognises HotelRunner, Elektraweb, SiteMinder, Cloudbeds, SynXis, D-EDGE and others from your home page. The agent picks dates, guests and a room and goes as far as the guest-details form, typing nothing into it. If booking happens on another domain, it says plainly that this part couldn’t be measured.
HotelsSupported today
Hotel structured data
Measures Hotel / LodgingBusiness data, amenities, room and price offers, cancellation terms and the booking entry with hotel-specific checks.
HotelsRoadmap
Google hotel prices
A readiness check for Google’s hotel prices and free booking links. On the roadmap.
B2BSupported today
B2B stores
Detects log-in-to-see-price walls, wholesale / case / dealer-only constraints, minimum order quantities and “price on request” offers, and says plainly why an agent can’t buy.
B2BSupported today
Request-a-quote flow
Measures whether an agent can find the product or service, reach the quote form and understand what it asks for. The form is never submitted.
ServicesSupported today
Service businesses
Measures service pages, contact details, pricing / “how it works” information and ProfessionalService / LocalBusiness data with service-specific checks.
ServicesSupported today
Contact and booking forms
The agent reaches your contact or quote form and reads its fields; it never fills or submits it. It also reports whether the form only loads via script.
ServicesRoadmap
Google Business Profile
Whether your name, address and phone match your website. On the roadmap.
ManufacturersSupported today
Manufacturer and brand product data
Measures GTIN / MPN / brand confidence, attribute coverage per category and a readiness score per product; opens missing attributes as tasks without inventing values.
ManufacturersRoadmap
Dealer and where-to-buy pages
Whether an agent can get from a manufacturer’s site to the right seller and stock. On the roadmap.
Data & feedsSupported today
Google Merchant Center
Reads your public product feed (or a feed URL or CSV you add) and compares price, stock and GTIN with your pages.
Data & feedsSupported today
OpenAI product feed
Generates and validates an OpenAI product feed from your catalog; you download it or share its stable URL. We don’t send it to OpenAI, and Turkey isn’t in the program yet.
Data & feedsSupported today
Server logs
Upload nginx / Apache logs, or on paid plans send them continuously with Cloudflare Logpush or a small script. AI bots are grouped by purpose and checked against their published IP ranges, and AI-referred visitors and the crawl-to-refer ratio are shown.
Data & feedsRoadmap
Akeneo PIM
Carrying PIM enrichment through to agent channels. On the roadmap; today we measure from your product pages and feed.
Data & feedsBuilt · not open yet
GA4 and Search Console
Tracking AI-referred traffic and traffic changes through a read-only Google connection. Built; it opens in the panel once we switch on the Google connection.
Store platformBuilt · not open yet
Write-back to Shopify and WooCommerce
Applies suggested fixes, such as structured data, to your store one at a time: only with your approval, with a diff to review and one-click undo. Built; it opens once we switch on the store apps. Until then, you apply the ready-made code yourself.
Agent protocolSupported today
UCP and agent discovery files
The free test checks your UCP profile, MCP / A2A / ai-catalog discovery files and Web Bot Auth directory, shown as information and not scored.
Agent protocolSupported today
WebMCP
The simulator lists WebMCP tools registered on the pages it visits and adds safety notes; it never calls them.
Agent protocolSupported today
Specoria scan API and MCP server
Run the free test from your own scripts or CI (POST /v1/scan, plus a GitHub Action) and read your projects’ tasks from an AI assistant over MCP (POST /v1/mcp), with an API key from your panel.
Agent protocolRoadmap
Hosted MCP endpoint
An endpoint hosted for you that agents can query directly for stock, shipping and return terms. On the roadmap.
Agent protocolRoadmap
Agentic Commerce Protocol
Commerce endpoints for checkout inside the agent, with the user’s approval. Today we only check for a hint of it; implementation is on the roadmap.
MarketplaceBuilt · not open yet
Trendyol
Connects read-only with your own Trendyol seller API key; matches your listings to your catalog by barcode and stock code and shows price and stock conflicts, GTIN problems, title quality and empty required attributes. Marketplace pages are never scraped and your listings are never changed. Built; opens once our legal review is complete.
MarketplaceRoadmap
Hepsiburada
Read-only price and stock comparison through your Hepsiburada seller (Listing) API; opens once our integrator registration is approved. On the roadmap.
PaymentsRoadmap
iyzico
Turning installment and payment options into offers agents can read. On the roadmap.
PaymentsRoadmap
Stripe
Shared payment tokens for agent-initiated payments. On the roadmap.
No matching integrations. Write to us and we’ll look into it together.
TEAMS
On your own or with a team: the same evidence, an answer for you
If you run the store on your own, the panel starts with quick wins that need no developer. With a team, leadership sees priorities, the catalog team sees missing fields and developers see ready-made code, all from the same measurement.
Store owner asks
“I run the store on my own and have no developer. Can I fix this myself?”
Specoria
No developer? The code is ready, with steps for your platform. Every task in the panel comes with ready-made code and steps for Shopify, WooCommerce, ikas, Ticimax, IdeaSoft or T-Soft, and the panel starts with quick wins that need no developer.
- CODEReady-made code for each task
- GUIDESteps for your platform
Each fix is re-measured with “Check” in the panel, so you know it worked.
Leadership asks
“Where is the agent channel losing us, and what will it take to fix?”
SpecoriaSAMPLE ANSWER · ILLUSTRATIVE
7 of the 12 sampled pages are missing a field agents check: GTIN, delivery time or return window. Three data gaps explain most of the findings, and each comes with ready-made fix code.
- AUDITDeep audit · product pages
- PANELTasks with fix code
On paid plans, a weekly “what changed” digest lands in your inbox every Monday.
Catalog team asks
“Which fields should we fix first?”
SpecoriaSAMPLE ANSWER · ILLUSTRATIVE
There are 1,240 critical missing fields. Priority order: GTIN (412), delivery time (388), return terms (290). Fix suggestions are ready, with the list of affected products.
- FEEDMerchant Center · 3,812 products
- WEBProduct pages and structured data
Every fix comes with ready-made code and is re-measured with “Check”.
Engineering asks
“How will agents connect to our systems, and how much load will that add?”
SpecoriaSAMPLE ANSWER · ILLUSTRATIVE
Specoria doesn’t need write access to your systems; we measure your store’s public face. Ready-made code (JSON-LD, robots.txt) is tailored to your platform. Our scan API and MCP server for your team are live, with an API key from your panel.
- CODEJSON-LD · robots.txt
- APIScan API · MCP
Every fix is re-measured with “Check” in the panel.
RESULTS
Methodology →
Not a visibility score. The decision the agent makes.
SAMPLE PROJECT · HOME APPLIANCES · 10 WEEKS · REPRESENTATIVE
Share of agent recommendations that include you
4% →31%
W1W2W3W4W5W6W7W8W9W10
CRITICAL MISSING FIELDS
GTIN, shipping and return fields completed across the entire catalog.
TOP REASON FOR BEING DROPPED
The remaining drop-offs now come from your commercial decisions, not your data.
FIXES VERIFIED BY RE-MEASURING
Ready-made code shipped, then re-measured: every fix counts only once an agent can read it.
Representative sample project: the chart and figures are not a real customer result.
TRUST
We don’t sell rankings. We build readiness.
No manipulation
No hidden text, no prompt injection, no fake reviews. Only accurate, verifiable data.
Read-only to start
We don’t need write access to your systems. Optional write-back to Shopify and WooCommerce, only with your approval and one-click undo, is coming.
GDPR
No customer data is processed, only catalog and policy data. Details are in our privacy notice.
Open standards
schema.org and GS1 today, and our scan API and MCP server for your team are live. A hosted MCP endpoint and ACP for your store are on the roadmap. Infrastructure that doesn’t lock you into a single platform.
FAQ
Questions store owners ask: access, results, pricing
The free test and the free deep audit + panel cost nothing and need no card. Starter’s first month is $1, then the list price on our pricing page. You cancel in one click from the panel, with no minimum term; details are in cancellation and refunds.
Specoria tests the path an agent takes through your store, not only whether an AI mentions you: 27 checks on your public pages, agents in a real browser, “couldn’t verify” never counted as a pass, and fix code with steps for your platform. What each kind of tool does, without naming anyone: Specoria and other tools.
No. The changes happen in structured data, feeds, robots.txt and policy pages. Your storefront stays exactly as it is.
Visibility asks ChatGPT, Perplexity and Gemini; simulations on paid plans run Claude-, ChatGPT- and Gemini-based agents in a real browser; Copilot gets a readiness estimate from the same checks. The panel expands as models are updated.
None. We only read your public pages, robots.txt, sitemap and product feed. Your team applies the ready-made code, or your expert does it with you on Growth.
You see the free test result within seconds. Leave your email and your deep audit report and panel arrive within minutes. Each fix is verified by re-measuring.
No. Spectora is home inspection software; Specoria measures how ready online stores are for AI shopping agents. More on who we are.
Agents are already shopping.
Free deep audit + panel: leave your store’s address and your email. Within minutes we send the PDF, made in a real browser, and open your panel with tasks and ready-made code.
Free test: enter your store’s address and see your result from 27 checks in seconds, right here. No email needed.
No sign-up · no email · no card
What the test reads, all in one request:
- robots.txt and agent permissions
- Your homepage and business details
- A product page, if one can be found
- Product data, the way an agent reads it
Scan limit reached
Each browser gets 1 free scan per week. Open your result below, or use your free panel to re-test anytime and get your full deep audit.
View your result Open my free panel →Free deep audit + panel: we’ll email you the PDF with fix recommendations and a link to your panel within minutes.
No credit card · No access to your systems · PDF in minutes · Starter’s first month is $1
We only use your email to send your deep audit and your panel link.
The test only reads your public pages; results are stored to prepare your report.