Evaluating GEO vendors - what questions actually reveal if they know what they're doing?

I’ve evaluated 20+ vendors for clients. Here’s my framework.

Technical understanding test:

Results validation:

The killer question:

“Show me a page that went from zero AI citations to regular citations. Walk me through exactly what you changed.”

If they can’t answer this with specifics, walk away.

Technical questions that separate experts from pretenders:

RAG understanding:

Ask: “How do you optimize content for retrieval-augmented generation?”

Good answer includes:

• Content structure for passage extraction
• Semantic coherence
• Information density
• How AI systems query knowledge bases

Vector search knowledge:

Ask: “How do vector embeddings affect our visibility?”

Good answer includes:

• Mathematical representations of content
• Semantic similarity in high-dimensional space
• Content clustering effects
• How AI matches queries to content

Crawl access expertise:

Ask: “How do you ensure AI systems can access our content?”

Good answer includes:

• Specific crawler user agents (GPTBot, PerplexityBot, ClaudeBot)
• robots.txt configuration
• JavaScript rendering issues
• Server-side rendering importance

If they can’t go deep on these topics, they’re just rebranding SEO.

How to validate case study claims:

What to ask for:

How to verify:

1. Ask for actual screenshots of AI responses
2. Request the specific prompts used
3. Ask to see monitoring tool dashboards
4. Check if they can replicate results live

Warning signs in case studies:

• Only showing ranking improvements (not AI citations)
• Vague metrics like “improved visibility”
• Single platform results only
• No timeline or date information
• Can’t show the actual monitoring setup

The best vendors will show you Am I Cited or similar dashboards with real data.

Platform-specific knowledge is essential.

Ask: “How do you optimize differently for each AI platform?”

Good answers include:

If they say “we optimize for all of them the same way,” that’s a red flag.

Follow-up questions:

• “What percentage of Perplexity citations come from Reddit?” (Answer: ~6.6%)
• “How does ChatGPT’s web browsing mode differ from its base knowledge?”
• “Why do only 11% of domains get cited by both ChatGPT and Perplexity?”

Real experts know these nuances. Pretenders give generic answers.

Measurement approach questions that matter:

Ask: “How do you measure GEO success?”

Red flag answers:

• “We track keyword rankings”
• “We measure organic traffic”
• “We look at backlinks”

Good answers:

• Citation Rate (pages cited / pages tracked)
• Response Inclusion Rate (prompts with brand / total prompts)
• Citation Share (your citations / total citations for query)
• Platform-specific visibility scores

Ask: “What’s your measurement methodology?”

Good answers include:

• Query panel definition (25-30 priority prompts)
• Multi-platform tracking
• Repeat-prompt sampling (AI responses vary)
• Competitive benchmarking

The measurement question separates vendors who understand GEO is fundamentally different from those just rebranding SEO.

Incredible insights. Here’s my evaluation checklist.

Technical Understanding (Score 1-5):

Results Validation (Score 1-5):

Red Flag Checklist:

• Guarantees rankings
• Focuses only on keywords
• Can’t explain RAG or vector search
• Treats all platforms the same
• No AI-specific case studies
• Only traditional SEO metrics

The interview structure:

1. Ask technical questions first (15 min)
2. Request case study walkthrough (20 min)
3. Discuss measurement approach (15 min)
4. Ask about platform-specific strategies (10 min)

Scoring:

• Score 4-5 on all areas: Proceed to proposal
• Score 3 on any area: Probe deeper
• Score 1-2 on any area: Disqualify

Thanks everyone for helping build this framework!