Intent First: How Conversational AI Needs a New Approach
- The modern customer has just one need that matters: Getting the thing they wont when they want it.
- Instead, intent-first architecture uses a lightweight language model to parse the query for intent and context, before delivering to the most relevant content sources (documents, APIs, peopel).
- Organizations are deploying LLM-powered search applications at a record pace,while a fundamental architectural issue is setting most up for failure.
The modern customer has just one need that matters: Getting the thing they wont when they want it. The old standard RAG model embed+retrieve+LLM misunderstands intent, overloads context and misses freshness, repeatedly sending customers down the wrong paths.
Instead, intent-first architecture uses a lightweight language model to parse the query for intent and context, before delivering to the most relevant content sources (documents, APIs, peopel).
Enterprise AI is a speeding train headed for a cliff. Organizations are deploying LLM-powered search applications at a record pace,while a fundamental architectural issue is setting most up for failure.
A recent Coveo study revealed that 72% of enterprise search queries fail to deliver meaningful results on the first attempt, while Gartner also predicts that the majority of conversational AI deployments have been falling short of enterprise expectations.
The problem isn’t the underlying models. It’s the architecture around them.
After designing and running live AI-driven customer interaction platforms at scale, serving millions of customer and citizen users at some of the world’s largest telecommunications and healthcare organizations, I’ve come to see a pattern. It’s the difference between prosperous AI-powered interaction deployments and multi-million-dollar failures.
It’s a cloud-native architecture pattern that I call Intent-First. And it’s reshaping the way enterprises build AI-powered experiences.
The $36 pillion problem
Table of Contents
Gartner projects the global conversational AI market will balloon to $36 billion by 2032. Enterprises are scrambling to get a slice. The demos are irresistible. Plug yoru LLM into your knowledge base, and suddenly it can answer customer questions in natural language.Magic.
Than production happens.
A major telecommunications provider I work with rolled out a RAG system with the expectation of driving down the support call rate. Instead, the rate increased. Callers tried AI-powered search, were provided incorrect answers with a high degree of confidence and called customer support angrier than before.
This pattern is repeated over and over. In healthcare, customer-facing AI assistants are providing patients with formulary data that’s outdated by weeks or months.Financial services chatbots are spitting out answers from both retail and institutional product content. Retailers are seeing discontinued products surface in product searches.
The issue isn’t a failure of AI technology. It’s a failure of architecture
Why standard RAG architectures fail
The standard RAG pattern – embedding the query,retrieving semanticall
Intent classification service
The classifier determines user intent before any retrieval occurs:
ALGORITHM: Intent Classification
INPUT: user_query (string)
OUTPUT: intent_result (object)
1. PREPROCESS query (normalize, expand contractions)
2. CLASSIFY using transformer model:
“`html
In healthcare deployments,the Intent-First pattern includes additional safeguards:
Detecting Customer Frustration with AI
Companies are increasingly using artificial intelligence to handle customer service, but a common problem is misinterpreting frustrated customers.A new approach, dubbed “Intent-First,” prioritizes understanding why a customer is contacting support before attempting to answer their question.This can dramatically improve customer experience and reduce escalations to human agents.
The core idea is to identify keywords that signal customer frustration. These aren’t necessarily related to the topic at hand, but rather to the customer’s emotional state. For example, detecting words like “terrible,” “hate,” or phrases like “still waiting” should promptly flag a conversation for human intervention.
Here are some examples of frustration detection keywords:
- Anger: “terrible,” “worst,” “hate,” “ridiculous”
- Time: “hours,” “days,” “still waiting”
- Failure: “useless,” “no help,” “doesn’t work”
- Escalation: “speak to human,” “real person,” “manager”
When these keywords are detected, the system should bypass automated search and immediately route the customer to a human support agent.
Cross-industry applications
The Intent-First pattern isn’t limited to one industry. It’s valuable wherever companies use conversational AI to manage diverse content. Here’s how it can apply in a few key sectors:
| Industry | Intent categories | Key benefit |
| Telecommunications | Sales, Support, Billing, Account, Retention | Prevents “cancel” misclassification |
| healthcare | Clinical, Co |