AI search index, issue 1
Issue 1 of our AI search index: we read 319 answers from 4 AI assistants to questions that named no agent, and 97.2% of them named at least one agent or team.
Almost every answer named an agent
We asked 4 AI assistants 3 questions per market, the way someone looking for an agent might ask who to hire, in 11 rounds of questions. None of the questions named an agent.
Of the 319 answers we read, 97.2% named at least one agent or team, and 657 different agents and teams were named in all. What stands out: an assistant asked for an agent almost always answers with names, so the question for you is whether yours is one of them.
Agent sites were cited as often as the portals
We counted the kinds of sources the answers cited. An answer can cite several sources, so these are shares of all the sources cited, not of the answers. Directory and review sites made up 25.6%, agent, team and local firm sites 23.5%, and listing portals 22.6%.
The rest were other sources at 10.5%, brokerage sites at 6.8%, social media at 5.3%, news at 3.7% and government sites at 2.1%. What stands out: sites that agents, teams and local firms run themselves were cited slightly more often than the listing portals, so your own site is one of the sources the assistants read.
| Share | |
|---|---|
| Directories and review sites | 25.6% |
| Agent, team and local firm sites | 23.5% |
| Listing portals | 22.6% |
| Other sources | 10.5% |
| Brokerage sites | 6.8% |
| Social media | 5.3% |
| News | 3.7% |
| Government sites | 2.1% |
What we released
In the month this issue covers, we put live, among others, briefing files for assistants, structured data on every page, website statistics that keep AI reads and AI visits apart, and the prompt battery that produced this issue.