Every answer an engine gives starts with reading. Before your business can be found, recommended, or quoted, a machine has to get in the door and make sense of what it finds. Two checks, sixty seconds total.

Type yoursite.com/robots.txt in a browser. That file tells crawlers what they may read. Look for the word Disallow followed by a slash under any user agent with AI in its name: GPTBot, ClaudeBot, PerplexityBot. A Disallow there means you built a site and then told the engines to leave. It happens more than anyone expects, usually from a security plugin or a host setting nobody reviewed.

Open an engine that browses, Perplexity or ChatGPT with search, and ask: "Read this page and tell me what this business does, where it operates, and what it costs: [your homepage]." What comes back is what the machine can actually see. Text trapped inside images is invisible. Menus that exist only as scripts can vanish. If the summary misses things a human sees on your page, the machine could not read them.

Remove any AI Disallow you did not choose on purpose. Make sure every important word on the page is real text, not a picture of text. Put the plain answer near the top, because engines lift the passage that answers.

Run both checks today and write down what the machine missed. That list is your next week of fixes, and every one of them compounds. Tomorrow: the quiet checklist an engine runs before it puts your name in an answer.

This post contains an affiliate link. If you start a trial through it, I may earn a commission at no additional cost to you.