The agent gathered a serviceable overview of Drafted's core offering, target markets, and pricing model from a single homepage fetch, but the site left critical details sparse—notably exact pricing, tier names, export-format mappings, and competitive positioning. The agent correctly flagged multiple confusions (conflicting stats, rendering glitches, staging-domain links) that reflect real site quality issues rather than navigation failure.

• ›The agent fetched only the homepage and did not explore linked pages (e.g., /learn/use-cases/prepare-for-architect, role-specific onboarding flows), so the answer is constrained to what the landing page explicitly stated. The site published no pricing details, competitor comparison, or technical docs on the homepage.
• ›The homepage contained inconsistent statistics (two different user counts, plan counts, and render counts across stat blocks) and rendering artifacts (duplicated button text like 'Get StartedStart Creating'), which the agent correctly identified as data quality issues, not navigation gaps.
• ›Calls to action pointed to staging.drafted.ai rather than the production domain (drafted.ai), creating ambiguity about whether the agent was inspecting the live product or a development version—a red flag for site navigability and credibility.
• ›The agent acknowledged it did not search for competitor information and therefore could not substantiate Drafted's claimed differentiators (speed, end-to-end workflow, schematic+furnished pairing) against alternatives, limiting the completeness of the 'how it differs' section.
• ›Pricing was deliberately deferred to in-app onboarding ('free to start' with no tier names or prices listed), forcing any prospective user to sign up to learn costs—a pattern that hides information from agent-readable discovery.