# Know what AI assistants can find out about your product I'm Erik. I check what AI assistants can read on your website or API, where the information disagrees, and what they say about your product. You get a written report explaining what I found, what to fix first, and how to check the changes. For Shopify stores, I offer a focused check of product information and the shopping journey before payment. We work in writing. I reply within one business day. [Scope and pricing](/services). Technical agent-readiness of turva.dev: 100/100 and Level 5, Agent-Native, on isitagentready.com. Measured 2026-09-06. Business ID 3600281-7, registered in Finland, based in Tampere and run by Erik Rekola. ## One product. Three different answers. - Visible product page: A real price and lead time - Structured product data: €0.00 · InStock - Product API: Empty price · Marked purchasable An assistant may find a different price or availability depending on which source it reads. What to fix: make the published product facts agree. How to check: compare the product page, structured data and API with the same product record. This example uses an invented business and invented readings. The sample's scanner does not score this issue. [See the finding and its evidence](/samples/audit-report#f1) ## Choose the check you need - [Shopify agent storefront check](/shopify-agent-storefront-check). €999. Do your selected products show the same price and availability across the shopping interfaces your store exposes? I check one store, one market and up to three named product and variant pairs. Delivered as one package of four written deliverables within 48 hours of the agreed written kickoff, followed by one retest of up to two corrected items within 14 days of that package. - [Website and API agent-readiness audit](/agent-readiness-audit). €4,300. Find out what automated clients can access and what selected AI assistants say about your product. I combine a technical scan, manual review and a recorded set of AI questions, and one re-scan within 30 days of the report is included. Delivered in two weeks. Prices exclude VAT. Each service can be bought on its own. Implementation, ongoing advisory, agent operations and MCP server design are scoped separately. [All services and prices](/services). ## See what the work looks like Sample reports. See how I explain a finding, show the evidence and describe the fix. Both reports use fictional businesses. [Sample audit report](/samples/audit-report) and [sample Shopify report](/samples/shopify-agent-storefront-check). Research. Read what I measured across company websites and AI assistants. Each study includes its methods, dates and limits. These are research observations, not client results. [Website agent-readiness study](/blog/website-agent-readiness-567-sites), [What AI assistants call an agent-readiness audit](/blog/what-ai-assistants-call-an-agent-readiness-audit) and [Thirty days after the brief](/blog/thirty-days-after-the-brief). My own site. I use turva.dev to put this work into practice. You can inspect the source code and the published checks. [Public scanner](https://isitagentready.com/) and [read the source](https://github.com/erekola/turva-worker). Scanner: isitagentready.com (third party, Cloudflare). Discoverability, Content Accessibility, Bot Access Control, and API, Auth, MCP and A2A Discovery: 100/100. Commerce: 100/100. Verified 100/100, Level 5, Agent-Native. I also publish the site's security checks. They measure different things from the agent-readiness scan, and like it they are measurable rather than asserted. Measured 2026-09-06. - Hardenize: all 24 categories passed. https://www.hardenize.com/report/turva.dev - Internet.nl website test: 98/100. https://internet.nl/site/turva.dev/ - Internet.nl email test: 95/100. https://internet.nl/mail/turva.dev/ ## How we work Tell me what you need to understand. Send the URL and your question. I confirm the scope, price, required access and start date in writing. Receive the report. I explain what I checked, what I found and which changes deserve attention first. Make and check the changes. Your team can use the correction instructions, or you can buy implementation separately. The follow-up check follows the terms of your chosen service. You work directly with me. I'm an independent consultant based in Tampere, Finland. [About Erik and turva.dev](/company). ## Frequently asked **What does agent-readiness mean?** It means how well an AI agent can find and use the information your website or API publishes. This work examines your product's public information and interfaces. **Is that the same as AI visibility?** No. Technical checks show what an agent can access. AI visibility checks record what selected assistants actually say about your business. The audit covers both and reports them separately. **Do you guarantee mentions or a score?** No. I report what I observe and how it was checked. AI answers and scanner results can change. **Can our own team make the fixes?** Yes. The report includes correction instructions and ways to check the result. Optional implementation and its access requirements are explained on the service pages. **Do you need our passwords?** The audit does not require production credentials. Any access needed for separately purchased implementation is agreed before that work starts. ## Tell me what you want to understand Send your website, API or Shopify store URL and the question you have. If you're unsure which service fits, describe the problem in your own words. [Request an audit](/contact). Email: . I'll reply within one business day and explain the next step. ## Markdown views You are reading the markdown view of this page. Every page on this site has one, and there are two ways to it: the page URL with .md appended, which is what llms.txt v2 asks for, or the page URL itself with Accept: text/markdown. Both return this same document, at a fraction of the token cost of the HTML.