Alignment

What Alignment measures

Alignment compares what your site says you are against what AI says you are when it's grounded with third-party content. If your hero says "fast-onboarding CRM for agencies" but live AI grounded responses describe you as "developer tooling," your Alignment is low.

Critical: Alignment is NOT a sentiment score. A critical review that accurately describes your positioning (even if negative) scores higher on Alignment than a glowing review that mischaracterises what you do. Sentiment-style risk lives in the separate Risk modifier.

The alignment probe

We extract your brand's stated positioning from your site (hero, About page, key feature pages) using LLM analysis. Then we probe ChatGPT, Gemini, and Perplexity with grounded queries about your category and parse their responses for how they describe you. The score measures the semantic overlap between the two.

• person_search Site-side positioning extraction — what does your own copy say you are, who you serve, what you do?
• forum Grounded AI synthesis — what do the third-party sources AI cites say you are?
• align_horizontal_center Semantic match score — how much do the two narratives agree on positioning, audience, and category?

How to improve Alignment

Alignment is a forcing function for honesty. Lift it by closing the gap between site copy and reality.

• edit_note Audit your hero / About / pricing for accuracy. If reviews say you're hard to set up but your site says "5-minute onboarding," fix the onboarding (or own the imperfection in your copy).
• forum Engage in third-party narrative. Respond to reviews, contribute to forum threads about your category, write a Wikipedia article that matches your positioning. The third-party narrative isn't fixed — it's shaped by your participation.
• campaign Re-state positioning in syndicated content. Bylines, podcast appearances, conference talks. AI weights "the founder said this" highly when grounding.
• assignment_late Don't manipulate sentiment — manipulate clarity. Trying to suppress negative reviews backfires (Risk goes up). Better: be honest about what you're not, so AI describes you accurately rather than wrong.

Frequently asked questions

• Is Alignment a sentiment score?expand_more
  No — Alignment is a positioning-match score, not a sentiment score. A negative review that accurately describes your brand's positioning (even if critical) scores higher on Alignment than a glowing review that mischaracterises what you do. We do NOT penalise brands for having negative reviews if those reviews describe them accurately. Sentiment-style risks live in the separate Risk modifier.
• How do I improve Alignment?expand_more
  Alignment lifts when your brand site copy matches what third parties say about you — and vice versa. Make sure your About page, hero, and product descriptions match what reviewers, journalists, and forum posters say you do. If reviews say "buggy" but your site says "reliable," fix the bugs (or own the imperfection in your copy). Alignment is a forcing function for honesty.
• Why does my Alignment score change between audits even when nothing changed?expand_more
  Alignment depends on which third-party content surfaces in live AI grounding. New reviews, news cycles, or forum threads can shift the third-party narrative day-to-day. The 90-day drift re-audit (free with every paid audit) helps separate genuine alignment shifts from grounding-source noise.