The problem this solves
Comparison questions are the highest-intent moment in AI-assisted buying: someone asking an assistant to weigh you against a rival is hours from a shortlist decision. The sources cited in those answers effectively write your sales narrative for you - and today they are probably review aggregators, forum threads, and competitor-owned pages. Few brands even know which sources speak for them in comparisons, because auditing every X-versus-Y permutation by hand across multiple engines is a project nobody staffs.
Winning these citations is a construction problem with known physics: engines quote pages that answer the comparison directly, structure claims as extractable facts, and read as balanced enough to trust. Most vendor comparison pages fail on the last point - they read as advertising, and the engines route around them. What is missing is a plan grounded in who wins each specific comparison today and what the winning pages do right.
How the mission runs
- Enumerate the battleground. The agent maps every comparison question that matters: you versus each rival, rival versus rival, and the best-of shortlist prompts where comparisons hide. SERP data confirms which phrasings buyers actually search, so the question set reflects real demand instead of guesswork.
- Capture who wins today. AI Visibility runs each comparison question through the live engines and records the cited sources, the framing of every recommendation, and where you currently stand. This produces the baseline: which comparisons you win, which you lose, and exactly who speaks for you.
- Decode the winning pages. For each comparison, the agent inspects the cited pages and identifies what earned the citation: direct answers, structured comparison tables, dated claims, visible evenhandedness. These observations become the spec every new page must meet to compete for the same slot.
- Design the page architecture. Each planned page gets a blueprint: the verdict framing, the table structure, which honest competitor strengths to concede for credibility, and the schema markup the page needs. Conceding real strengths is deliberate - balance is what makes a vendor page quotable.
- Lay out the plan in Notion. The full program lands in Notion: a database of comparison targets with current citation holders, the page blueprint for each, and a priority order based on deal impact and winnability. It is a build plan your content team can start executing tomorrow.
The prompt
This is the exact objective the agent receives. Swap the obvious placeholders for your own domain, segment or channel and run it as-is from the console, Slack, or the API.
What comes back
A Notion workspace containing the comparison battleground: every X-versus-Y question mapped with its current citation winners and captured answers, a per-page blueprint specifying structure, framing, and markup, and a prioritized build order. The evidence behind every judgment - full answer captures and cited-page analyses - stays linked, so writers building each page can see exactly what they are trying to displace.
Make it yours
- Start with the three comparisons that appear most often in your lost-deal notes and expand the map after those first pages ship.
- Include the comparison prompts where you are absent entirely - category shortlist questions - since earning any mention there usually precedes winning head-to-heads.
- Ask the agent to also draft the two highest-priority pages from their blueprints, staged for your review alongside the plan itself.
Frequently asked questions
Why do engines prefer third-party comparison sources?
Cited sources skew toward pages that read as evaluative rather than promotional, and third parties clear that bar by default. Vendor pages can compete by adopting the same physics - direct verdicts, structured facts, honest concessions - and the plan is built around what the currently cited pages demonstrably do.
How current is the citation baseline?
Every comparison answer is captured live when the mission runs, timestamped and stored. Citations in this space shift as engines refresh their sources, so the plan records its observation date, and re-running the mission after pages ship shows whether they have started displacing the incumbents.
Should comparison pages really admit competitor strengths?
Selective, truthful concessions are what separate quotable pages from brochures. The blueprint recommends conceding points you genuinely lose while framing the dimensions where you win, because engines and buyers both discount pages that claim a clean sweep. Your team keeps final say over every concession.