All termsGlossary

What is entity extraction?

Entity extraction is the task of picking out specific details from what a customer says or writes, such as names, dates, amounts, addresses and order numbers, and labelling each one.

Why it matters

Knowing that a customer wants to change a booking is not enough to change it. The system also needs the booking reference and the new date. Entity extraction finds those details in free text or speech, so customers can explain things in their own words and are not forced through a form. It also saves agents from retyping details into other systems.

How it works

A model reads the message and marks the spans of text that belong to known types, such as date, product or postcode. Each value is then put into a standard form, so “next Friday” becomes a calendar date and a spoken number becomes digits. Values in strict formats, such as order numbers, are often checked against a pattern or looked up to confirm they exist.

Example

A customer emails, “Hi, order 48213 arrived yesterday but the blue kettle is cracked. Please send a replacement to 14 Mill Lane.” Entity extraction returns the order number, the delivery date, the product and the address as separate labelled values. The support system uses them to open the right order and prefill a replacement request for an agent to confirm.

Get Started

Let’s build AI agents, together.

Talk with us to see the top three use cases for AI agents at your business, and how quickly you could go live.

Live in 7 days · Works inside the tools your team already uses · Hands off with full context