An intent taxonomy is the structured, organized set of all customer intents a support operation handles, arranged so each conversation maps to one clear goal.
It is the vocabulary of customer goals. A good taxonomy is mutually exclusive (intents do not overlap) and collectively exhaustive (every real conversation has a home). It often nests: a broad area like billing contains specific intents like duplicate charge, refund request, and update payment method. The shape follows the business: an e-commerce taxonomy fills with where is my order, address change, and subscription pause; a B2B software one with login issues, plan changes, and API errors.
Why it matters: the taxonomy is the foundation everything else stands on. Classification, automation, reporting, and coverage measurement all reference it. A vague or incomplete taxonomy means vague automation and unmeasurable progress. A precise one makes the whole operation legible.
The Aide point of view: at Aide, the agentic AI platform for customer experience, the intent taxonomy is not hand-authored from guesswork. It is auto-discovered from real conversations and structured into a three-level Customer Intent Map, then maintained as the operation evolves. A clean taxonomy lets each intent be scoped, tested, and governed on its own rather than bundled into one opaque workflow. It also gives the team a shared vocabulary for why customers get in touch, one that stays current as the business changes instead of decaying behind the automation.
Frequently asked questions
- What is the difference between an intent taxonomy and an intent map?
- A taxonomy is the structured list of intents. A Customer Intent Map is Aide's auto-discovered, three-level realization of one, built from real conversations.
- How many intents should an intent taxonomy have?
- There is no fixed number. It should be exhaustive enough that every real conversation maps to one intent, and precise enough that no two intents overlap.