The AI recommendation gap scorecard ranks the keywords where AI doesn't recommend you, so you fix the ones most likely to pay off first.
Short answer: score each gap from 1 to 5 on four factors – search demand, commercial intent, competitive weakness and strategic fit – add the scores, and sort. Work from the top. Once you have a gap list it is quick to fill in, and it stops a team from spending months on a head term that three big competitors already own.
How do you use the scorecard?
List your gaps. Start with every keyword or prompt where AI assistants name competitors but not you. If you don't have that list yet, run a keyword-level AI recommendation audit first.
Score each factor 1–5 using the scales below. Agree the scales as a team before you start, so two people would give the same score.
Add the scores to get a priority score out of 20.
Sort from highest to lowest and assign the top five to your next content sprint.
Re-score monthly. Competitors and AI answers change, so a gap that was open can close.
What goes in the scorecard?
Keyword | Search demand (1–5) | Commercial intent (1–5) | Competitive weakness (1–5) | Strategic fit (1–5) | Priority score |
|---|---|---|---|---|---|
[Keyword 1] | |||||
[Keyword 2] |
Search demand
AI assistants don't publish prompt volumes, so use Google monthly search volume as a stand-in. People tend to ask assistants similar questions to the ones they search. Treat it as a rough guide, not a precise measure.
5: over 10,000 searches a month
4: 1,001–10,000
3: 501–1,000
2: 101–500
1: 0–100
Commercial intent
How close the question is to a buying decision. Gaps on decision-stage questions are where disappearing from AI answers costs the most.
5: transactional ("[category] pricing", "buy [product]")
4: comparison ("best [category] for [use case]", "[competitor] alternatives")
3: how-to with product context ("how to track AI visibility")
2: broad informational ("what is generative AI")
1: unrelated or navigational
Competitive weakness
How easy the recommendation is to win. A high score means the field is open. Run the prompt a few times in each assistant before scoring, because one answer can mislead.
5: AI names no one, or gives a vague answer
4: only weak or off-target brands are named
3: one strong competitor is named
2: two strong competitors
1: three or more strong, established competitors
This is the factor teams skip most. Chasing "best project management software" against the category giants can take months with nothing to show for it, while several mid-intent questions with weak competition stay open.
Strategic fit
Whether winning this keyword brings the customers you want.
5: directly about your core product
4: a main use case or problem your buyers have
3: a secondary feature or adjacent topic
2: a broad industry topic
1: tangential
What does a filled-in scorecard look like?
A hypothetical example for a company selling an AI content platform:
Keyword | Demand | Intent | Weakness | Fit | Priority score |
|---|---|---|---|---|---|
AI content generation platform | 4 | 4 | 2 | 5 | 15 |
how to improve brand visibility with AI | 3 | 3 | 5 | 5 | 16 |
what is marketing AI | 5 | 2 | 3 | 2 | 12 |
"How to improve brand visibility with AI" comes out on top. It has less demand than the other two, but the field is open and it fits the product exactly. "What is marketing AI" has the most demand and the least value, which is the trap the scorecard is meant to catch.
How can you adjust the scorecard?
Weight what matters most. If you need quick wins, double competitive weakness. In the example the scores become 17, 21 and 15, so the open-field keyword pulls further ahead. If you need leads fast, double commercial intent instead.
Add an effort score. Subtract 1 for a short update to an existing page and up to 3 for a new data-heavy guide, so cheap wins rise.
Add the cited sources. Note which pages the assistants cite for each gap. If they're review sites you can't influence, lower the weakness score.
For a fuller version of the research step, see how to map every keyword where AI isn't recommending your brand.
What should you do with the top of the list?
Write one page per top gap that answers the question in its first two sentences, then backs it up with specifics, fair comparisons and dated prices. Publish, wait two to four weeks, re-run the prompts and re-score. Platforms that combine tracking and drafting, such as blogawesome (from $49 a month), can shorten the step from gap to draft. The writer quoted on our writers page went from about 2 hours per article to 1. The scorecard works just as well with a spreadsheet.
To get your first gap list, run the free AI visibility audit. It takes no signup and about 30 seconds, and shows where AI models recommend competitors instead of you.