It handles long, careful work
A single project means large research sets, long fact sheets, and strict style rules. Claude works well with long inputs and detailed instructions.
Uses Claude, checked by people
How we use Claude to build pages AI search can cite
We build mainly on Claude, Anthropic's AI model, through Anthropic's commercial API. Claude does the research, mapping, drafting, and fact-checking. People set the strategy and approve every page. Here is exactly how a project runs.
The six stages of every project
- 01
Research
What Claude does
Claude research agents gather the questions buyers ask AI assistants in your market, the searches that show a Google AI Overview, and the sources those answers cite.
What people do
We choose the market, the services, and the locations that matter to your business.
Output
A research set of real questions, AI answers, and cited competitors
- 02
Entity map
What Claude does
Claude turns the research into an entity map of your business, services, places, and proof. It also drafts a fact sheet from your website and documents.
What people do
You approve every fact. Prices, licenses, service areas, and claims come from you, not from the model.
Output
An entity map and a signed-off fact sheet
- 03
Draft
What Claude does
Claude writes answer-first pages using only the approved fact sheet, following our structure, schema, and style rules.
What people do
We set the page plan, the internal links, and the rules each draft must follow.
Output
Structured drafts with schema markup
- 04
Fact-check
What Claude does
A second Claude pass checks every claim in each draft against the fact sheet and flags anything it cannot match.
What people do
We resolve every flag: fix the claim, find the source, or cut it.
Output
Drafts with zero open flags
- 05
Review
What Claude does
Nothing. This stage is people only.
What people do
An SEO specialist reads, edits, and approves each page. You sign off before anything goes live.
Output
Human-approved pages, ready to publish
- 06
Track
What Claude does
Claude runs regular checks of AI answers and search results for your brand and summarizes what changed.
What people do
We review the numbers, decide what to fix next, and report to you in plain language.
Output
Citation checks, AI traffic, calls, and form requests in one report
Why we build on Claude
Business terms that protect your data
We use Claude through Anthropic's commercial API. Anthropic says it does not use inputs or outputs from its commercial products to train its models by default.
Source: Anthropic Privacy Center (opens in a new tab)We still pick the right tool
Claude is our main model, not our only one. When another model does a specific task better, we use it, and the same human review applies.
Our quality and data rules
AI makes the work fast. These rules keep it accurate and safe. Our AI policy has the full detail.
- Every page is read, edited, and approved by a person.
- Every fact on a page comes from a fact sheet you approved.
- We never invent reviews, credentials, prices, or statistics.
- We send AI models only the data a task needs.
- We never put your confidential data into consumer chat apps.
- We add an AI disclosure to your site when you, your platform, or the law requires one.
See what this system would build for your business
Send a short form. A person reads it, and we aim to reply within one business day.
Claude and Anthropic are trademarks of Anthropic, PBC. Generative SEO is independent and is not affiliated with, sponsored by, or endorsed by Anthropic.