AI · Brand · Independent. Case study by Eugen Stoicov, Senior Product Designer in Vienna.
Brands spend on Google, then ask ChatGPT who to buy from. I made a way to read the answer, including when it is unflattering.
Queries ChatGPT, Claude and Gemini about a brand, then scores visibility, sentiment and who else gets named. Often the models prefer someone else.
People now ask a chatbot who to hire, who to buy from, which clinic to book. The answer is often a competitor, or a shrug. Marketing teams could not see that without sitting down and prompting three models by hand.
ChatGPT, Claude and Gemini disagree in ways you cannot average away. Scores have to be honest. The site says so: no inflated numbers. Each live run costs real API calls, so a stranger on the homepage cannot dump an unbounded report. Free is 5 analyses and 1 brand. Pro is $49. Enterprise $199.
Lead with a URL, then the account, because the report is the expensive part. The artefact is a 0-100 score, a side-by-side of ChatGPT / Claude / Gemini, a competitor rank, and a punch list of pages to fix. I refused a dashboard that hides the raw reply.
Live at botsandbrands.xyz. The customers page still says coming soon. The product does not. I still wince at some of the answers. That is the point.
Lead with a URL, not a deck. stripe.com is the example on the field because everyone already knows what a yes should look like.
Visibility out of 100, then ChatGPT, Claude and Gemini side by side. Recommends, mixed, or not. The interesting case is when they disagree.
Competitor rank, then a punch list: add the certification page, publish pricing, write the case studies. I still find that last one funny, given this site.
Five analyses, one brand, $0. Pro at $49 if the score is ugly enough to keep watching. The live run still asks for an account, because each query costs real model calls.
Focus: 0→1 product, AI visibility, Competitive intel, Identity.