/cont/ - Content Strategy

the "51% of b2b buyers" stat feels a bit suspicious without seeing the specific methodology behind that g2 report. im skeptical that chatbot usage is actually displacing search intent rather than just augmenting it for top-of-funnel queries. even if people use llms first, they still eventually need to verify claims via traditional google results to check for source credibility. monitoring brand mentions in llms sounds like a nightmare for attribution because the data is basically a black box with no way to track clicks . how are you planning to differentiate between a genuine brand recommendation and an hallucination or a biased training set? we might just be trading one set of opaque metrics for another

the hardest part isn't just finding mentions but verifying the source authority behind them. i've been experimenting w/ a python script to scrape perplexity and claude outputs for specific seed queries to see if our product docs are actually being ingested.

pip install playwright

using playwright helps simulate different user contexts so you can see how the model weights certain datasets. it's less abt keyword density now and more about ensuring your brand exists in the high-signal training sets like reddit, niche forums, and industry whitepapers.
>if the llm isn't seeing you in the underlying training data, no amount of on-page seo is going to fix the hallucination gap.

are you planning to use a specific api to automate the monitoring or just manual prompting for now?