Forget the hype. Forget the dashboards and “AI SEO kits.” Before you spend a dollar, you need to know one thing: how AI engines describe your brand today.
This isn’t about buying tools. It’s about running a quick, no-nonsense test to see your current state. You can do it in 30 minutes with nothing more than ChatGPT, Gemini, or Claude — and a spreadsheet.
Step 1: Run Three Types of Prompts
AI engines don’t work like Google. They answer in contexts. To see how you show up, you need to test three prompt types:
1. Customer Needs/Interests Prompts (Tier 1)
These mimic what your buyers would actually ask.
- “I’m a procurement manager looking for eco-friendly suppliers in the packaging industry.”
- “What’s the best IT support provider for mid-sized law firms?”
- “Which marketing analytics platforms are most cost-effective for startups?”
👉 Why it matters: This shows whether you even surface when real buying questions are asked.
2. Brand Understanding Prompts (Tier 2)
These test whether the AI understands you at all.
- “Who are the main competitors to [your brand]?”
- “Describe [your brand] in one paragraph.”
- “What is [your brand] best known for?”
👉 Why it matters: If the AI misunderstands or misses your differentiators, customers will too.
3. Product/Service Prompts (Tier 3)
These test at the offering level — where sales happen.
- “What’s the best [product/service] under $X?”
- “Which [brand] [product] is right for [specific use case]?”
- “Compare [product A] vs. [product B].”
👉 Why it matters: Even if your brand is mentioned, you may lose at the product level if competitors are positioned better.
Step 2: Record What You See
Use a simple grid (Excel, Google Sheets, or our Prompt Benchmarking Grid if you want structure). For each prompt, log:
- Brand Presence (✅/❌) – did you appear?
- Competitors Present – who else showed up?
- Description Accuracy (1–5) – did AI get you right?
- Recommendation Relevance (1–5) – was the answer useful to a buyer?
- Notes – what was missing, wrong, or surprising?
Step 3: Compare Across People & Models
- Run the same prompts in ChatGPT, Gemini, Claude.
- Have 2–3 people run them separately.
👉 AI outputs vary — and those differences are where you’ll find blind spots.
What You’ll Know in 30 Minutes
- Visibility: Are you in the conversation where it matters?
- Positioning: Does AI describe you correctly — or not at all?
- Competitive Map: Who’s winning in AI answers you’re losing?
This isn’t a kit. It’s not a product. It’s a reality check.
Next Step: Make It Measurable
Repeat quarterly using the same prompts. That gives you a trend line of:
- Visibility share vs. competitors
- Accuracy of brand positioning
- Movement of competitor presence
Without this baseline, you’re just guessing. With it, you know if GEO is an urgent issue, a competitive gap, or just noise.
Bottom Line
You don’t need dashboards, agencies, or “AI kits” to start. You just need 30 minutes and the right prompts. Because if you don’t know how AI engines are answering your customers’ questions, your competitors will be happy to answer on your behalf.
