Who do you trust?

We often talk about Garbage in-Garbage out when it comes to AI. The quality of your outcome depends on the quality of the data it has to work with.

The team at Semrush, who help organisations become cited sources in AI, published their latest AI visibility index. Primarily targeted at brands who want to show up in Google AI, there are some interesting results of what sources are cited the most when used with different types of requests.

Would you take your professional and business advice from Reddit?

Well for much of the summer they were the highest cited source used by ChatGPT, followed by Wikipedia. Into the Autumn, they were clearly unprioritised.

This leads to a bigger challenge, which is, you can only cite what is publicly available. If an AI engine can't see it, they can't cite it.

If you are using publicly available data, then you are using the sources that are available. They often aren't great and rely on blog posts and opinions which may be unchecked. The sentiment can be biased too.

Method: 2,500 prompts across five industry verticals, with equal
distribution among subcategories within each vertical. Branded queries were excluded to ensure unbiased results and no prompts overlapped across categories. Tests were run using ChatGPT with web search on, done bi-weekly with some daily refreshes.

This data will change as there is a greater availability of information.

SOURCE
Top 10 Sources: https://ai-visibility-index.semrush.com/#sourcesTop10
Full report: https://lp.semrush.com/rs/519-IIY-869/images/The-2025-AI-Visibility-Index-Study-Semrush-Enterprise-compressed.pdf

BESCI AI OPINION


We often talk about the Knowledge Stack, where organisations curate the knowledge they want their AI agents and models to use for better advice and results.

If you feel that the AI slop is really 'internet slop', based on the average of the opinions that are open and available to be cited, then consider curating your own knowledge stack.

Many of our clients are using our Irrational Change knowledge to inform their change, transformation and leadership results. In the future we expect to see a marketplace for knowledge to make it easier to choose what and whose voice you want your teams to hear.

Previous
Previous

AI Companionship

Next
Next

90% of Code will be written by AI