Glossary
What is entity extraction?
Definition
Entity extraction is the process of identifying and pulling specific pieces of information out of text or speech. These entities include names, dates, phone numbers, addresses, amounts, and other structured details. Entity extraction turns unstructured language into data a system can act on.
01How entity extraction works
The system scans an utterance and labels spans of text as belonging to entity types such as person, location, time, or product. Approaches range from rule and pattern based methods to machine learning models that learn from annotated examples. In dialogue systems this is often called slot filling, where the assistant collects each required piece of information.
02Entity extraction in voice systems
When a caller says 'schedule John Smith for July tenth at two,' entity extraction identifies the name, date, and time. These values populate the fields needed to complete a booking, look up an account, or route a call. Because callers give details in unpredictable order and formats, robust extraction normalizes them into consistent values.
03Why it matters
Extraction accuracy directly affects whether the right appointment, contact, or record is created. Ambiguous inputs, misheard numbers, or unusual names can cause errors, so systems often confirm critical details back to the caller. Combining extraction with validation reduces mistakes on high-stakes data like phone numbers and dates.
Frequently asked questions
Is entity extraction the same as intent detection?
No. Intent detection identifies the caller's goal, while entity extraction captures the specific details, such as dates or names, needed to fulfill that goal.
What is slot filling?
Slot filling is entity extraction applied to dialogue, where the assistant gathers each required value, or slot, before completing a task like booking an appointment.
See also
Related terms
Ahoya is an AI receptionist that answers every call 24/7.
Start free