Legal AI Search Index — planned research
This page describes a proposed research project. Collection has not been completed and no index scores, rankings, sample results or customer outcomes are published here. The initial 50-prompt panel is a starting instrument, not evidence of search demand.
The questions we intend to investigate
Which firms appear for different legal buyer situations? Which domains are cited? How often do directories, firm websites or publications supply linked sources? How do answers differ by engine and change over time? We will distinguish observed citations from claims about hidden source influence.
Sampling plan
Begin with the downloadable 50-prompt panel, covering immigration, employment, personal injury, family and business-law buyer situations in Toronto. The location is a pilot choice, not a claim of market representativeness. Expand to hundreds of prompts across selected cities and regions only after fixing the protocol and versioning additions.
Collection plan
Run ChatGPT, Gemini, Perplexity and Google AI experiences under documented conditions. Record Google AI Overviews and AI Mode separately where available. Predefine paraphrases and repeated runs; preserve timestamps, raw answers, named firms, source domains, citation URLs and failures. Record when no AI answer appears.
Publication requirements
Before releasing findings, publish the collection dates, sample definition, product modes, complete prompt panel, repeat counts, exclusions, denominators and limitations. Review citation coding and firm-name matching. Release usable source records where permitted and explain any redactions. Do not publish an index based only on favorable examples.
What future findings will distinguish
Report appearance frequency, citation frequency, description accuracy and relevance separately by engine, practice area and prompt group. A share of mentions is a statistic for that panel, not an estimate of total market share or buyer demand. No observations are available yet.
Prior work and limits
The GEO study provides a research starting point for content experiments. Google’s published guidance supports attention to accessible, useful pages and ordinary SEO foundations. Neither source validates an unpublished Twenty Five Labs index or a guaranteed gain for a particular firm.
The first 50 buyer-intent prompts
Download the 50-prompt pilot panel (CSV). All prompts are unbranded discovery questions. They have not been run or validated as demand research. Use only services and jurisdictions relevant to the firm under study; retain the original pilot as a distinct version if you adapt it.
Proposed output fields and denominators are documented in our measurement methodology.
Sources and scope
Platform documentation supports the technical context. Our page plans and measurement protocol are proposed methods. They do not demonstrate customer results or reveal private ranking systems.
Google: AI features and your websiteOpenAI: crawler controlsAggarwal et al.: GEO — Generative Engine Optimization (2024 revision)