Answer engines are excellent at narrative context — what a company says about itself, recent coverage, who competes with it. They are unreliable at identity, which is exactly what your CRM needs.
The split that works
Use registers for the spine: legal name, registration number, status, registered address, directors, sector codes. Use Perplexity for the story: positioning, recent news, likely priorities, people quoted in the press.
Never let the narrative layer overwrite a registry field. Store them separately, with the source recorded, and show the register value as the authoritative one.
Where it goes wrong
Trading names, holding structures and same-name entities in different markets trip up web-sourced answers. A confident summary of the wrong legal entity is worse than no summary.
Recency is also uneven: a story from three years ago can outrank last month's filing in a web index.
Being visible to answer engines
If you want your own company to be described accurately by these tools, publish structured, crawlable facts: a clear about page, consistent legal name and address, and schema markup.
That is the same discipline behind AI visibility work — the engines can only repeat what they can find and parse.
Frequently asked questions
Can Perplexity replace a company data provider?
No. It can summarise what the web says, but it does not guarantee a stable identifier, complete coverage or consistent fields across a market — which is what list building and CRM hygiene require.
How do I make my company easier for AI search to describe?
Keep legal name, address and description consistent everywhere, publish structured data, and make key facts available on crawlable pages rather than only in images or PDFs.
