A property management platform that AI crawlers could not read, and whose marketing sold a different product

Context

What we found

• The original build served non-rendering crawlers an empty page
• The marketing site, its llms.txt and its schema described facility and building management, while the product's own data model was built for rental portfolios, owners and agencies
• Page-level structured data existed on only two page types
• The default language setting pointed search engines to the wrong version for the main market
• A wrong product name visible on the live contact page, and a framework version with a published security advisory

What we did

• Defined the target segments and a first wedge: owners and long-term rental agencies, with corporates and investors second
• Rewrote the site-wide schema as a connected graph covering the organisation, website, apps and service, plus breadcrumbs and collection pages per template
• Repositioned metadata, llms.txt, page descriptions and the FAQ answers that feed FAQPage markup
• Fixed the default locale and alternate language references, and added search verification
• Built six location and seasonal pages, claiming a physical office only where one exists
• Built a content plan and an article engine for the team

Result

• 14 files changed in the client's repository, passing strict type checks with zero errors and delivered deploy-ready
• The first location page approved by the owner and used as the template for five more

Services used

More work