Skip to content

DABLOCK DATA DESK

How the DABLOCK AI Visibility Index is measured

Answer. A fixed panel of 16 category buyer prompts is submitted to 3 AI engines (openai, perplexity, gemini). For each answer we record whether a tracked brand is named. Share of answer = the share of panel prompts naming the brand, averaged across engines with equal weights. Re-measured weekly. Paid placement never affects a score.

Current release

Rules

Prompt panel

Panel version 3. A share of answer is a share of this list, so the list is frozen between releases and every measurement records the version that produced it. We do not compute a change across two different panel versions — see the archive, where such rows read «new panel» instead of a movement that did not happen.

  1. best crypto exchange for beginners
  2. best crypto exchange with the lowest trading fees
  3. safest crypto wallet for long term storage
  4. best hardware wallet for bitcoin and ethereum
  5. best self-custody wallet for solana
  6. best decentralized exchange to swap tokens
  7. best defi platform to earn yield on stablecoins
  8. best way to stake ethereum without running a node
  9. best crypto portfolio tracker app
  10. best on-chain analytics tool to track wallets and flag risky addresses
  11. best site to check crypto prices and market data
  12. best smart contract audit firm for a new defi protocol
  13. best crypto tax software for filing capital gains
  14. best nft marketplace to buy and sell nfts
  15. best way to buy crypto with a credit card
  16. best ethereum layer 2 network for cheap transactions

Engines we run, and the one we dropped

This release runs 3 engines: ChatGPT, Perplexity, Gemini. A third, Google Gemini, was measured and then excluded, and saying so here matters more than the tidiness of a two-engine table. Run through our relay it returned zero mentions for every brand across the whole panel — a result that is not a finding about the market but a failure of the measurement. Publishing it would have put a column of zeros next to real numbers; dropping it silently would have been worse, because the engine count is part of what the score means.

So the rule is written into the code rather than left to judgement: a build in which any engine returns zero across the entire panel refuses to publish at all. Weights are re-normalised across the engines that did answer, never filled with an assumed value. Gemini returns to the index when the relay is fixed and it produces a measurement we can defend — and its return will be a new panel version, not a quiet edit.

Ownership disclosure

DABLOCK is published by VECTORY, an AI-visibility company. This is disclosed here, in the footer of every page and in /humans.txt, because a measurement is only useful if you know who ran it. VECTORY sells measurement and advisory services; it does not sell positions in this index.

Reuse

Data is published under CC BY 4.0. Cite as: DABLOCK AI Visibility Index — Crypto & Web3, 2026-08-31. dablock.ai

Machine copies: /api/aiv.json, /aiv.csv, /index.md, /llms.txt.