How do I benchmark us against competitors in Copilot results?

To benchmark “us vs competitors” in Copilot / AI answer results, treat it like a share-of-answer + quality-of-answer study, not a classic SEO rank check. Copilot is dynamic, query-sensitive, and often blends web results, citations, and model-generated summaries.

1) Define the competitor set and query set

Start by choosing:

• Your top 3–10 competitors
• 20–100 target queries
  • Brand queries: “Best [category] software”
  • Problem queries: “How to do X”
  • Comparison queries: “[You] vs [competitor]”
  • Category queries: “[category] tools for [segment]”
  • Bottom-funnel queries: “pricing”, “alternatives”, “reviews”

Group queries by intent:

• Awareness
• Consideration
• Decision

2) Decide what you’re measuring

For Copilot-style answers, useful metrics are:

Visibility metrics

• Mention rate: how often your brand appears in the answer
• Competitor mention rate
• Top-3 inclusion rate: whether you’re named among leading options
• Citation/share of citations: whether the answer cites your site vs competitors
• Link inclusion rate: whether a source link points to you

Sentiment / positioning metrics

• Positive / neutral / negative framing
• Feature parity: which features are mentioned for each brand
• Recommendation strength: are you explicitly recommended, or just listed?
• Differentiation clarity: do answers explain why you’re better/worse?

Accuracy metrics

• Claim accuracy: are facts about your product correct?
• Message consistency: do the same talking points show up repeatedly?

3) Run a repeatable test

Because AI outputs vary, test each query multiple times:

• Use the same prompt wording
• Run each query 3–10 times
• Test in:
  • Signed-in vs signed-out state
  • Different geographies, if relevant
  • Different devices/browsers if you suspect personalization

Capture for each run:

• Query
• Date/time
• Prompt
• Response text
• Citations/links
• Whether you or competitors are mentioned
• Sentiment and recommendation status

4) Score the results

Create a simple rubric, for example:

Per query/run scoring

• 0 = not mentioned
• 1 = mentioned but not recommended
• 2 = positively framed or cited
• 3 = explicitly recommended / preferred

Then compute:

• Average score per query
• Average score by intent
• Average score vs each competitor
• Mention share:
  • Your mentions / total competitor mentions

5) Compare on key “decision” dimensions

For each competitor, assess whether Copilot answers emphasize:

• Price
• Ease of use
• Enterprise/security
• Integrations
• Support
• AI features
• Market traction / reviews
• Best fit by segment

This reveals whether competitors “own” certain topics in AI answers.

6) Build a dashboard

A basic spreadsheet is enough at first. Columns:

• Query
• Intent
• Your mention? (Y/N)
• Competitor A mention? (Y/N)
• Competitor B mention? (Y/N)
• You cited? (Y/N)
• Competitor cited? (Y/N)
• Recommendation score
• Sentiment
• Notes / excerpt

If you want a more advanced setup, use:

• BrightEdge / Conductor / Semrush / Ahrefs for broader visibility tracking
• A custom script or manual process for Copilot outputs
• A BI dashboard to trend results over time

7) Benchmark against “AI answer quality,” not just mentions

Sometimes the winner is not the most-mentioned brand, but the one with:

• More accurate descriptions
• Better citations
• Stronger category association
• More frequent inclusion in “best for X” contexts

So compare:

• Presence
• Positioning
• Proof
• Persuasion

8) Watch for content and technical drivers

If competitors dominate Copilot answers, look at why:

• Their pages are clearer and better structured
• They have stronger third-party coverage
• Their brand/entity signals are stronger
• They have more authoritative backlinks and reviews
• Their content directly answers the query better

9) Use the findings to improve

Typical actions:

• Create comparison pages: “You vs Competitor”
• Add concise FAQs and definitions
• Improve schema markup
• Strengthen thought leadership and third-party mentions
• Publish pages that answer “best X for Y” use cases
• Make claims easy for AI to extract and cite

10) A practical benchmark template

For each query, record:

• Query
• Intent
• Your brand mentioned? Y/N
• Competitor(s) mentioned? Y/N
• You cited? Y/N
• Competitor cited? Y/N
• Answer sentiment toward you: positive/neutral/negative
• Explicit recommendation: you / competitor / mixed / none
• Notes

Then summarize:

• Visibility share
• Citation share
• Recommendation share
• Sentiment share

---

If you want, I can give you:

1. a ready-to-use spreadsheet template, or
2. a step-by-step Copilot benchmarking framework for your specific industry and competitors.