# Revision: Cheat Sheet and Self-Check — AI Visibility and LLM Brand Discovery

Source: https://www.geekswithgeeks.com/en/ai-visibility/wrap-revision

> Review the concepts, techniques and ethics of AI visibility from the whole course.

## Cheat sheet

**Concept**: assistants give one answer with few brands; visibility = mention, citation, recommendation, accuracy, sentiment; knowledge comes from training memory and live retrieval. **Influence**: consistent entity facts, independent authority, answer-first useful pages. **Technical**: decide crawler access per bot (robots.txt), server-rendered content, JSON-LD that matches the page, sitemap, llms.txt as an experiment. **Measurement**: fixed prompt set across intents and languages, repeated runs, mention rate with a range, share of voice, citation/domain analysis, referrers and server logs. **Playbooks**: pricing, comparisons, how-tos, FAQs, original data; honest reviews; Hindi and local content. **Risks and ethics**: fix wrong facts at their sources, address negative themes by improving the product, never use hidden prompts or fake reviews. **Operate**: monthly loop, change log, honest scorecard, no guarantees.

## Questions interviewers ask

Be ready to explain: how assistants learn about a brand (memory versus retrieval), what GEO adds to SEO, why client-side rendering can hurt, how you would measure visibility fairly given non-deterministic answers, how to respond when an assistant states a wrong fact, and which tactics are unethical.

**Quiz:** Your pricing page content is injected by JavaScript. Which check best reveals the risk?

- [ ] Change the font
- [ ] Count your social followers
- [x] Fetch the page with curl and see if the price text is in the raw HTML
- [ ] Rename the domain

*Answer:* Fetch the page with curl and see if the price text is in the raw HTML. If the text is absent from the raw HTML, crawlers that do not run JavaScript will miss it.

**Quiz:** Brand mentioned in 6 of 10 answers. How should you report this?

- [x] 60% with a wide range, and collect more runs
- [ ] Exactly 60%, nothing more is needed
- [ ] 100% visibility
- [ ] 0% because samples are small

*Answer:* 60% with a wide range, and collect more runs. With only ten answers the plausible range is wide, so show it and add runs before drawing conclusions.

**Quiz:** An agency promises "guaranteed #1 mention in all AI assistants within 30 days". What is the right reaction?

- [ ] Believe it because it is confident
- [ ] Sign immediately
- [ ] Pay in advance for the full year
- [x] Be sceptical: nobody controls assistant outputs, so no guarantee can be honest

*Answer:* Be sceptical: nobody controls assistant outputs, so no guarantee can be honest. Outputs vary and change, so honest providers talk about improving the odds and measuring, not guarantees.