Deep scan asks the real answer engines (Perplexity + ChatGPT) your buyers' questions across repeated runs and reports your citation share with a 95% confidence interval. Validate lift runs a significance test between a baseline and a post-fix scan so you can prove a change is real, not noise.