Map the starting scope
A short intake captures the product, priority entry points, buyer decisions, audiences, conditions, and pages you want considered. I help define the boundary before the audit.
How it works
A guided audit. A workspace your team can use.
You bring the product context. I audit what your site communicates, explain the findings through a UX and product design lens, and guide your team through Product Atlas so you can resolve the facts together.
Book a free 15-minute assessment ↗A typical initial engagement
From first conversation to a reviewed implementation.
Some clients start with one priority page. Others have several connected entry points that need to be understood together. The assessment sets a useful scope for the audit, Atlas work, implementation, and testing.
Extract, audit, and compare
I collect public page wording, metadata, relevant structured data, and source URLs. I manually compare the entry points with the pages they depend on, and include an agreed competitor comparison when it helps clarify positioning. I record conflicts, missing conditions, and questions a machine extraction alone cannot settle.
Walk through the findings and Atlas
We review the page map and consequential issues together. I bring a UX and product design lens to the offer and its flow, then show how evidence, owner decisions, and audience modes work in Product Atlas.
Your team resolves facts
You review, edit, qualify, or leave each claim open in Atlas. I guide the decisions as needed. Audience modes can change the explanation and next step while sharing approved facts, and you can export working files at any point.
Request the final review
When your team is done, send the current project export for my final review. I check unresolved decisions, source links, conditions, and consistency, then prepare the agreed implementation files. The public sample uses an email handoff with an exported JSON file; shared server saving is planned for the client installation.
Publish, test, and hand off
Your team or I implement the agreed changes according to scope. We repeat relevant parser, page, and answer checks after publication, label what each result establishes, and hand off the source record and selected files.
The handoff
Formats follow the destination.
Product Atlas can preview canonical JSON, page HTML, JSON-LD, Markdown, XML, CSV, and YAML. The engagement produces the formats named in the agreed implementation, not a pile of files with nowhere to go.
If your industry has a specific schema, feed, taxonomy, or import template, we can map approved fields to that destination after checking what it accepts. Unsupported fields remain explicit rather than being forced into the wrong vocabulary.
SOURCE BEFORE OUTPUT
One approved record.
Claims retain evidence, conditions, owner decisions, page relationships, and review status. Structured formats are generated from that record, then validated in the destination that will consume them.
Explore the sample Atlas ↗Schema.org and ingestion
Design for the routes that matter. Verify the ones we can see.
Schema.org vocabulary helps express supported entities and relationships in structured data. It is one route among several. Search systems decide what to crawl, index, display, and send to models, so each route receives an appropriate check.
RouteOur treatmentVerification boundary
Raw HTMLPlace essential meaning in the server response and inspect it without browser execution.Directly testable
Visible textKeep claims and qualifications readable in page text, including when scripts are removed.Directly testable
Structured dataAlign eligible entities to supported Schema.org types and test the published markup.Parser testable
Linked files and feedsProvide named JSON, Markdown, or feed outputs where a consumer will follow them.Destination dependent
Selected chunksRepeat crucial conditions near claims so a partial excerpt retains its meaning.Sampled, not controlled
Browser agentsUse clear headings, links, labels, actions, and visible state for a navigable page.Scenario testable
This work improves the information a system can encounter. It does not promise that a search provider will index it, display a rich result, or use it in an AI answer.