How a long-established building-products supplier became a working research site for capturing business knowledge, restructuring legacy information and connecting paper processes to modern software.
01
D.W. Ross Company was founded in Cleveland in 1921 and incorporated in 1972, and four generations of the Ross family have run it. What it knows about insulation, cold-storage panels, doors and windows, and the customers who buy them, lives mostly in people's experience and on paper. The website, last rebuilt around 2002, carried a small part of it.
How does a business keep its operating knowledge so that people, websites, software and AI systems can all work from facts someone can stand behind?
Four connected pieces of work since September 2026. Each shows its steps, where a person checks the result, and its status.
Their old site was static HTML, last rebuilt around 2002, and the domain had since passed to a broker. We rebuilt it from the archive and from conversations with the owners. Every page now states its facts in a form search engines and AI can read: the business record, 29 services, and a plain-text file written for AI.
Seventeen working sessions with the owners and staff, recorded with AI meeting notes. A person corrects the notes against what was actually said. Then each statement becomes a dated ruling with a name on it, and the website is built from the rulings, never from the raw notes.
Staff send job photos from a phone with a one-line note. For each batch, an AI session decides which page the photos belong on, writes the captions, blurs faces and house numbers, and proposes the change. A person approves it before anything goes live.
Someone in the office photographs a paper vendor quote. The AI reads it but never sets a price. A person checks and corrects every field, then sends it, and an estimate appears in QuickBooks. The tool can create estimates and nothing else: no invoices, no payments, nothing sent to a customer.
None of this is measured in hours saved yet. When it is, we'll publish the number and how we got it.
03
Modernization started well before anything a customer could see. The business first had to decide which knowledge mattered, who could confirm it, how it should be structured, and which systems were allowed to use it. AI sped up parts of the work: reading meeting notes, sorting photos, reading a paper quote. Whether the result could be trusted depended every time on evidence, review and context.
The fieldwork produced a testable method and working systems. It isn't evidence of a general business outcome.
04
It feeds our research into Business Truth: a lasting, checkable record of a business that websites, search, AI answers, analytics and controlled software can work from. The research is ongoing. We're sorting out which parts generalize, which need a person every time, and which can become repeatable systems.
Before any of this was built, the discovery found the business described differently across its own pages, its Google listing, directories and AI answers. These findings were the starting point, not the story.
| Fact | Source | What it said | Record | State |
|---|---|---|---|---|
| What they sell | Google listing | Left out windows. | Windows, doors, flooring and insulated panels. | Confirmed · |
| Closing time | Business listings | Said they close at 4:30. | They close at 4:00. | Confirmed · |
| Install or supply | Business directory | Said they install windows. | They supply windows and work with qualified installers. | Confirmed · |
| What AI says | AI assistants | Nobody had checked. | What AI told customers about them was unknown. | Unknown · |
Real findings from D.W. Ross Company discovery, September 2026. Shown with permission.