AI Visibility
Index.
Cross-engine benchmarking of brand selection and citation frequency across ChatGPT, Perplexity, Gemini, Claude, Microsoft Copilot. Published under CC BY 4.0 — free to cite with attribution to BackTier.
Disclosure: this Index is an internal BackTier experiment, not an independent audit. BackTier is excluded from its own rankings. Scores are not independently verified; use them as directional, and read the methodology before citing. Journalists: see the Index press room and the media outreach page for story leads, quotes, and the press kit.
Prompts run 6 July 2026 to 24 July 2026 · Published 27 July 2026 · Next edition v2026.2 on 26 January 2027
AI Visibility Agencies
| # | Brand | AIV Score | ChatGPT | Perplexity | Gemini | Claude | Copilot | Citation Share |
|---|---|---|---|---|---|---|---|---|
| 1 | Profound | 71.5 | 74 | 72 | 70 | 68 | 73 | 11.8% |
| 2 | Otterly.AI | 64.1 | 63 | 68 | 60 | 62 | 67 | 7.9% |
| 3 | Peec AI | 58.7 | 59 | 61 | 55 | 57 | 61 | 5.2% |
| 4 | AthenaHQ | 55.3 | 55 | 57 | 52 | 54 | 58 | 4.6% |
Enterprise SEO Platforms
| # | Brand | AIV Score | ChatGPT | Perplexity | Gemini | Claude | Copilot | Citation Share |
|---|---|---|---|---|---|---|---|---|
| 1 | Semrush | 88.6 | 90 | 89 | 87 | 86 | 91 | 22.1% |
| 2 | Ahrefs | 85.4 | 87 | 86 | 84 | 83 | 87 | 19.7% |
| 3 | Moz | 70.2 | 71 | 72 | 69 | 68 | 71 | 9.3% |
| 4 | BrightEdge | 63.9 | 64 | 66 | 62 | 61 | 66 | 6.1% |
Answer Engines
| # | Brand | AIV Score | ChatGPT | Perplexity | Gemini | Claude | Copilot | Citation Share |
|---|---|---|---|---|---|---|---|---|
| 1 | Perplexity | 92.8 | 93 | 99 | 90 | 91 | 91 | 28.4% |
| 2 | You.com | 68.3 | 70 | 72 | 66 | 65 | 68 | 8.7% |
| 3 | Andi | 52.1 | 53 | 55 | 50 | 49 | 53 | 3.9% |
Legal Services (US)
| # | Brand | AIV Score | ChatGPT | Perplexity | Gemini | Claude | Copilot | Citation Share |
|---|---|---|---|---|---|---|---|---|
| 1 | Morgan & Morgan | 78.4 | 79 | 80 | 76 | 77 | 80 | 14.2% |
| 2 | LegalZoom | 74.6 | 76 | 75 | 73 | 72 | 77 | 11.6% |
| 3 | Rocket Lawyer | 66.2 | 67 | 68 | 64 | 64 | 68 | 7.3% |
Release schedule
Twice a year — published 26 January and 27 July. Prompts are run in the two weeks before each publication date, so the run window and the publication date are always stated separately. Every edition stays online at its own version number; nothing is silently overwritten.
| Version | Period covered | Prompts run | Published | Status | What changes |
|---|---|---|---|---|---|
| v2026.1 — H1 2026 | 1 January 2026 to 27 July 2026 | 6 July 2026 to 24 July 2026 | 27 July 2026 | published | First public edition. Establishes the four scoring dimensions and the five-engine panel. |
| v2026.2 — H2 2026 | 28 July 2026 to 31 December 2026 | 4 January 2027 to 15 January 2027 | 26 January 2027 | scheduled | Same categories, same weights, so scores are directly comparable to 2026.1. Movement columns added against the prior edition. |
| v2027.1 — H1 2027 | 1 January 2027 to 30 June 2027 | 5 July 2027 to 16 July 2027 | 27 July 2027 | scheduled | Category list reviewed before the run; any added or retired category is listed in the edition notes. |
How the AIV Score is calculated
The score is a weighted composite of four measured dimensions, scaled 0 to 100, where 100 is first-named and cited in every sampled answer on every engine. Weights are fixed for the life of a major version so editions inside it stay comparable — changing a weight forces a new major version and a restated prior edition.
| Dimension | Weight | What it measures |
|---|---|---|
| Selection rate | 40% | How often the brand appears at all in an answer to a category prompt, across every engine in the panel. |
| Citation rate | 25% | How often the answer links to or names the brand's own pages as the source, rather than mentioning it in passing. |
| Position | 20% | Where the brand falls in the answer — named first, listed among several, or mentioned as an also-ran. |
| Entity accuracy | 15% | Whether the engine describes the brand correctly: right category, right location, no merged or confused identity. |
Methodology, cadence, and limitations
- Edition
- v2026.1 — H1 2026.
- Period covered
- 1 January 2026 to 27 July 2026.
- Prompts run
- 6 July 2026 to 24 July 2026.
- Published
- 27 July 2026.
- Engines
- ChatGPT, Perplexity, Gemini, Claude, Microsoft Copilot.
- Coverage
- 14 brands across 4 categories.
- Scoring
- Selection rate 40%, Citation rate 25%, Position 20%, Entity accuracy 15% — scaled 0 to 100, where 100 is first-named and cited in every sampled answer on every engine.
- Update cadence
- Twice a year — published 26 January and 27 July. Next edition v2026.2 is scheduled for 26 January 2027, from prompts run 4 January 2027 to 15 January 2027.
- Licence
- CC BY 4.0. The full data is downloadable above as CSV and JSON.
What this index cannot support as a claim: it is produced by BackTier's own prompt panel and is not independently audited; generative engines are non-deterministic, so the same prompt can return different answers on different runs; a score measures how often a brand was selected and cited in this panel, not product quality; and BackTier is both a measured brand here and the publisher, so its own row should be read with that in mind. Claim tagging follows our provenance and confidence standard. Errors are corrected in public in the corrections log under our editorial policy.
How to cite
BackTier. (2026). AI Visibility Index — H1 2026 Results (v2026.1). Published 27 July 2026. Retrieved from https://backtier.com/ai-visibility-index
Licensed under Creative Commons Attribution 4.0 (CC BY 4.0).
Start here
Your brand is invisible to AI.
Let's fix that.
Before any implementation, we document how each major assistant currently describes your brand — and how often it names someone else.
Request an AI Visibility Baseline
Thirty minutes to see how the major assistants answer for your category, and where your brand sits in those answers.