Intent Recognition
Definition
Intent recognition is the classification task at the heart of NLU. When a user sends a message, the intent recognition model maps that message to one of the intents the system has been designed to handle. An intent represents a user goal — what they want to achieve — regardless of how they phrase it. A single intent (e.g., cancel subscription) may have dozens of training utterances covering all the ways users might express that goal. The quality of intent recognition directly determines how often the bot responds relevantly versus falling back to a generic error.
Why It Matters
Intent recognition is the first decision point in any chatbot interaction. If the bot misidentifies the intent, every subsequent step is wrong — wrong response, wrong data fetched, wrong action taken. High intent recognition accuracy is foundational to a chatbot that users trust and return to. It also determines escalation rates: when intent recognition fails, the conversation must fall back to a human.
How It Works
Intent recognition models are typically trained as multi-class classifiers. Each intent has a set of example utterances (training data). The model learns to assign new messages to the most likely intent class. Transformer-based models like BERT achieve high accuracy even with limited training examples. At runtime, the model returns a confidence score for each intent; the highest-confidence intent above a threshold is selected, or a fallback is triggered if no intent scores high enough.
Real-World Example
A user types 'how much does the pro plan cost?' The intent recognizer scores this message against all defined intents and assigns 0.94 confidence to 'pricing_inquiry'. The bot then retrieves the pricing information for the pro plan and responds with the relevant details.
Common Mistakes
- ✕Defining intents that are too similar, causing the model to confuse them — for example 'cancel subscription' vs 'pause subscription' need to be clearly differentiated in training data.
- ✕Under-training intents with fewer than 10-15 example utterances per intent.
- ✕Setting confidence thresholds too low, causing the bot to act on low-confidence matches instead of asking for clarification.
Related Terms
Natural Language Understanding (NLU)
Natural Language Understanding (NLU) is the AI capability that interprets the meaning behind human text or speech — identifying what the user wants (intent) and extracting key details (entities). NLU is the 'comprehension' layer of a chatbot, translating raw input into structured information the system can act on.
Entity Extraction
Entity extraction is the process of identifying and pulling specific pieces of information from a user's message — such as names, dates, order numbers, or locations. These extracted values (entities) fill in the details the chatbot needs to complete a task, working alongside intent recognition to fully understand the user's request.
Slot Filling
Slot filling is the dialogue management process of collecting all the required pieces of information (slots) needed to complete a task. The chatbot systematically asks for any missing slots — like date, time, or account number — until it has everything needed to fulfill the user's request.
Fallback Response
A fallback response is what a chatbot says when it cannot understand the user's message or find an appropriate answer. Instead of returning an error or going silent, the bot delivers a graceful fallback — acknowledging the limitation and offering alternatives like rephrasing, browsing the FAQ, or speaking to a human agent.
User Utterance
A user utterance is any message, phrase, or spoken input a user sends to a chatbot. It is the raw input that the NLU layer processes to determine intent and extract entities. Understanding the variety of utterances users produce for the same intent is essential for training accurate, robust chatbot models.
Ready to build your AI chatbot?
Put these concepts into practice with 99helpers — no code required.
Start free trial →