Record yourself doing a live AEO diagnosis. On camera, with screen sharing and audio narration, do the following:

• Open any AI answer engine and submit your compound-job prompt.
• Read the AI answer. Narrate what you see: which brands are mentioned, which are cited, and what you notice about the answer's structure.
• Pick one winner and one loser from the answer. Navigate to the winner's website and explain what their page does well. Name the gate or gates they're passing and point to the specific elements on the page (the billboard, the answer blocks, the page architecture) that make the engine trust them.
• Navigate to the loser's website and explain what's failing. Name the gate that's broken and point to the specific structural problem on the page.
• Prescribe one fix for the loser. Be specific: what would the page look like after?
• Place the fix in the correct SAGE phase.

Record and upload a single, unedited video, 5–8 minutes without slides or script.

What compound-job prompt did you use?

A strong compound-job prompt includes a product, a situation, and at least one constraint that forces the engine to reason rather than retrieve. The more specific your prompt, the richer your diagnosis will be.

Example: "What project management tool should a 30-person agency use if they run 15–20 concurrent client projects, need time tracking that integrates with their billing system, and their PMs need client-facing dashboards they can share without giving login access?"

Which AI answer engine did you use?

Name the two brands you diagnosed and which was the winner and loser.

Which gate is the primary failure point for the losing brand?

Options:

• Fetchable
• Chosen
• Extractable

Decide the single gate that best describes the losing brand's primary visibility problem. Your video should walk through the evidence for this in detail. This question is a quick reference for the evaluator.

Which SAGE phase does your proposed fix belong to?

Study this citation table, then record your read.

The image above is a real Top Citation Pages view from Profound. You are looking at this data as HubSpot. The table is filtered to Google AI Overviews, ChatGPT, Gemini, and Google AI Mode.

Each row shows: the page URL, its category (Other, Earned Media, Social), whether HubSpot is or is not mentioned on that page, the page's citation count with its recent change, and the page's share of total citations with its recent change. There are 988 total items in the full dataset. You are seeing the top results.

We want to hear you answer these questions, in order, as you move through the table:

• What are you looking at first, and why does the course say to start there?
• What does the mix of source types (the categories and domain types in the table) tell you about the citation supply chain for this topic?
• Where is HubSpot present and where is HubSpot absent? What does the pattern of "Mentioned" vs. "Not Mentioned" reveal when you cross-reference it with citation count and share?
• What's changing? Which pages are gaining or losing citation share, and what does that movement tell you about where the opportunity is headed?
• Based on everything you just read: what is your single most important next action? Be specific. What page would you build, what outreach would you do, what would you investigate next, and why does this data point you there?

Record a single, unedited screen recording, 2–4 minutes, with voiceover narration.

Read this scenario. Decide whether the colleague's recommendation is sound. If it's flawed, identify the specific conceptual error, explain why their logic breaks down even though parts of it seem right, and propose what you'd do instead using the frameworks from the course.

The Scenario

Your company sells HR software. You've been tracking AI answer visibility for three months. Your colleague pulls up the fan-out data and notices that when people ask about "employee onboarding software," the AI engines frequently transform the query to include words like "compliance," "checklist," and "document automation."

Your colleague proposes the following: "The data is telling us what the AI thinks onboarding means. We should build content pages targeting each of these transformed words. A page for 'onboarding compliance,' a page for 'onboarding checklists,' a page for 'document automation for new hires.' If the AI is expanding our topic into these words, we should have a page for each one. That's how we get retrieved."

Is this recommendation sound? If not, what specifically is wrong with it, and what would you recommend instead?

Write in your own voice. We're reading for clear thinking, not polished prose. Reference specific course concepts and terminology.

Did you use any AI tools to write your Conceptual Defense response?

Are you completing this certification on behalf of an agency? If so, enter the agency name (optional).

Record yourself doing a live AEO diagnosis. On camera, with screen sharing and audio narration, do the following:

• Open any AI answer engine and submit your compound-job prompt.
• Read the AI answer. Narrate what you see: which brands are mentioned, which are cited, and what you notice about the answer's structure.
• Pick one winner and one loser from the answer. Navigate to the winner's website and explain what their page does well. Name the gate or gates they're passing and point to the specific elements on the page (the billboard, the answer blocks, the page architecture) that make the engine trust them.
• Navigate to the loser's website and explain what's failing. Name the gate that's broken and point to the specific structural problem on the page.
• Prescribe one fix for the loser. Be specific: what would the page look like after?
• Place the fix in the correct SAGE phase.

Record and upload a single, unedited video, 5–8 minutes without slides or script.

What compound-job prompt did you use?

A strong compound-job prompt includes a product, a situation, and at least one constraint that forces the engine to reason rather than retrieve. The more specific your prompt, the richer your diagnosis will be.

Example: "What project management tool should a 30-person agency use if they run 15–20 concurrent client projects, need time tracking that integrates with their billing system, and their PMs need client-facing dashboards they can share without giving login access?"

Which AI answer engine did you use?

Name the two brands you diagnosed and which was the winner and loser.

Which gate is the primary failure point for the losing brand?

Options:

• Fetchable
• Chosen
• Extractable

Decide the single gate that best describes the losing brand's primary visibility problem. Your video should walk through the evidence for this in detail. This question is a quick reference for the evaluator.

Which SAGE phase does your proposed fix belong to?

Study this citation table, then record your read.

The image above is a real Top Citation Pages view from Profound. You are looking at this data as HubSpot. The table is filtered to Google AI Overviews, ChatGPT, Gemini, and Google AI Mode.

Each row shows: the page URL, its category (Other, Earned Media, Social), whether HubSpot is or is not mentioned on that page, the page's citation count with its recent change, and the page's share of total citations with its recent change. There are 988 total items in the full dataset. You are seeing the top results.

We want to hear you answer these questions, in order, as you move through the table:

• What are you looking at first, and why does the course say to start there?
• What does the mix of source types (the categories and domain types in the table) tell you about the citation supply chain for this topic?
• Where is HubSpot present and where is HubSpot absent? What does the pattern of "Mentioned" vs. "Not Mentioned" reveal when you cross-reference it with citation count and share?
• What's changing? Which pages are gaining or losing citation share, and what does that movement tell you about where the opportunity is headed?
• Based on everything you just read: what is your single most important next action? Be specific. What page would you build, what outreach would you do, what would you investigate next, and why does this data point you there?

Record a single, unedited screen recording, 2–4 minutes, with voiceover narration.

Read this scenario. Decide whether the colleague's recommendation is sound. If it's flawed, identify the specific conceptual error, explain why their logic breaks down even though parts of it seem right, and propose what you'd do instead using the frameworks from the course.

The Scenario

Your company sells HR software. You've been tracking AI answer visibility for three months. Your colleague pulls up the fan-out data and notices that when people ask about "employee onboarding software," the AI engines frequently transform the query to include words like "compliance," "checklist," and "document automation."

Your colleague proposes the following: "The data is telling us what the AI thinks onboarding means. We should build content pages targeting each of these transformed words. A page for 'onboarding compliance,' a page for 'onboarding checklists,' a page for 'document automation for new hires.' If the AI is expanding our topic into these words, we should have a page for each one. That's how we get retrieved."

Is this recommendation sound? If not, what specifically is wrong with it, and what would you recommend instead?

Write in your own voice. We're reading for clear thinking, not polished prose. Reference specific course concepts and terminology.

Did you use any AI tools to write your Conceptual Defense response?

Are you completing this certification on behalf of an agency? If so, enter the agency name (optional).