naia
Public methodology

How naia Index calculates the public ranking

The naia Index methodology was designed to make AI visibility comparable, auditable, and useful for GEO decision-making.

Today the public ranking covers the Brazilian market with two engines, five prompt families, and weekly updates. This page documents the logic behind the score, the main reading limits, and the public endpoints around the initiative.

Covered engines
2
Prompts per industry
5
Maximum score
100
Refresh cycle
Weekly
Methodology

What enters the calculation

Each public naia Index run queries ChatGPT and Gemini with active web search. The goal is to simulate the type of question consumers actually ask when they evaluate a market in Brazil.

The pipeline then extracts the cited brands, records each position in every answer, and consolidates the score into a single segment ranking. Weekly publishing keeps enough history to read momentum, stability, and decline.

01
Real questions
We simulate the questions consumers ask ChatGPT and Gemini when searching for products and services.
02
Direct queries
We send each question to the AIs with active web search, capturing responses in real time.
03
Brand extraction
We identify all mentioned brands, their ranking positions, and the context of each citation.
04
Visibility score
We calculate visibility across comparable prompts, considering brand presence and position in the responses.

Prompts used in the public collection

The list below shows the exact format of the prompts used today for the public default segment. Only the segment name changes across markets.

  1. 1.Qual o melhor plano de saúde no Brasil?
  2. 2.Qual o plano de saúde mais barato no Brasil?
  3. 3.Qual o plano de saúde com melhor custo-benefício no Brasil?
  4. 4.Qual o plano de saúde mais confiável no Brasil?
  5. 5.Qual o plano de saúde mais recomendado no Brasil?

How scoring works

The public score was built to reward appearance frequency and top ranked positions. The ceiling for a single run is 100 points.

  • Each run uses 2 engines x 5 prompts. If a brand ranks first in every query, it reaches the 100-point reference ceiling.
  • Position scoring is fixed: 1st = 10 points, 2nd = 8, 3rd = 6, 4th = 4, and 5th = 2.
  • Weekly change compares the current score with the previous week in the same segment, preserving history instead of overwriting older results.

Cadence and update windows

The public naia Index groups results by the current week, always anchored on Monday. When a new run is completed, the segment receives the week's date and the previous history remains available for delta calculation.

How to interpret it correctly

  • The ranking measures presence in evaluated engine answers, not the intrinsic quality of the product or service.
  • The current public coverage observes ChatGPT and Gemini. Other models may diverge in specific markets.
  • The prompts are tailored for Brazil and reflect generic purchase intent. Highly technical niches may require additional queries.
  • Brands that look tied to a human reader can still appear separated because of small differences in ranking position and recurrence.
Go beyond the public ranking

Boost your visibility across AI engines with naia

The naia Index is the public surface. Inside the platform, your brand gets a private GEO Panel Rank, a robust 9-dimension score, and agents that continuously work to improve your presence on ChatGPT, Gemini, Claude, Perplexity, and Grok.

Private GEO Panel Rank
Track position, share of voice, and weekly variation of your brand and competitors across every AI engine.
Robust 9-dimension score
Goes well beyond position: measures visibility, citations, sentiment, consistency, authority, schema, sources, and risk.
Agents that work for you
naia generates content, AI-optimized pages, and actionable recommendations that grow your presence across generative engines.
Create free accountSee plans and creditsStart with no card. First analysis is free.