Most directory owners who think about switching platforms stop at the same worry: what happens to everything I've built? The members who pay you, the listings you've featured, and the pages that already rank in Google. A plain CSV import keeps names and addresses and drops the rest.
Directories.ai has an import built for this. It reads the export files from four platforms and keeps the parts a switching directory can't afford to lose. This post walks through what it keeps, how to run it, and how to protect your search rankings on the way over.
Which platforms it reads
The import recognises exports from:
Column names differ between versions and settings of each platform, so the import doesn't rely on exact headers. It matches columns by lists of known names, ignoring case and punctuation, and detects which platform a file came from by its headers.
What it keeps that a normal import drops
For each row, as well as the business details, the import carries over:
It also uses the same duplicate matching, import history and undo as every other import, so a run you don't like can be rolled back.
Option 1: Import from the dashboard
Open your directory's Bulk import page, go to the CSV / spreadsheet tab and choose your file. If the file looks like one of the four platforms, a banner says so, for example "This looks like a Brilliant Directories export", with a Keep plans and old URLs switch.
With the switch on, the banner tells you what will happen before anything is written: how many members on how many plans become membership tiers, how many featured listings stay featured, and how many old page URLs are kept for redirects.
Option 2: Ask your AI to do it
If you manage your directory from Claude, ChatGPT or another AI connected over MCP, the same import is a tool called import_competitor_export.
It's careful by default. With dry_run on, which is the default, it writes nothing. It shows the detected platform, which columns matched, the plans it found and a sample of rows. Look at that first, then ask the assistant to run it for real. Active listings come in as drafts unless you ask for them to be published, and you can file every new listing under one category.
A single call takes the file's text for up to 5,000 rows. Bigger exports go in as batches of up to 500 rows that continue the same import job.
Keep your rankings: redirects
Your old listing pages may have years of search history behind them. If those addresses stop working after the move, the rankings go with them.
That's why the import keeps every old page URL. Once the listings are published, the get_migration_redirects tool lists each old address next to the listing's new one. It also returns a ready-made redirects file you can paste into a _redirects file on Netlify or Cloudflare Pages, or turn into rules on your own server.
These are permanent (301) redirects, which tell search engines a page has moved for good so its standing carries over. Your assistant can put the file in a pull request on your site's repository for you to review, or simply hand it to you.
A sensible order of work
Getting started
If you're weighing a move, create a directory and try a dry run with your own export. It writes nothing until you say so. The MCP developer reference covers connecting an AI if you'd rather have it do the work, and the setup page has the steps for each AI app.