The Future Of Language Understanding Involves More Intents, Not Less
A few months ago I coined the term “Intent Driven Design & Development” (IDD). The premise is that chatbot design and development should start with intents. This is the process of using existing customer utterances and conversations to create clusters of semantically similar sentences.
You can call intents what you like, similarity clusters, topics, intent documents, etc. The fact remains that customer intent must be the basis for any successful chatbot deployment, chatbot analysis, defining applicability to the customer journey, and workflow design.
If customer intents are used for defining the domain and design of the chatbot, clarity is established for defining use cases, CX goals and ultimately technology choices.
For successful chatbot deployment, analyse and define the chatbot applicability to the customer journey, and design the chatbot workflow based on customer intent.
The long-tail of NLU
It is important to address the long-tail of NLU, and ultimately the aim is to address all NLU requirements (including the long-tail) with as little as possible data.
The advantage of NLU’s in current chatbot development environments is the limited amount of data required for training per intent. Often 20 user utterances are enough per intent for training a NLU model.
Deployment teams focus their energy on the features of chatbot technology and metrics instead of focusing on the customer intent.
Intend Driven Design and Development (IDD) leads to more granular intents, and highly granular intents leads to addressing the long-tail of NLU. This approach also establishes a much deeper taxonomy with identifying sub-intents or nested intents.
This leads to a situation where the long-tail is easily addressed and solved for. The utterances from users which pertains to the long-tail is within your customer conversation transcripts. But discovering those niche utterance clusters is the challenge.
Intents & User Experience Are Linked
Gartner confirmed that a lack of focus on intents leads to poor user experience. Current customer conversations and utterances can be leveraged from current customer care interfaces like IVR, email, etc. These utterances/conversations can be used for creating intent clusters which are semantically similar.
The popular and dominant current approach to chatbot projects is to start with technology and compiling a comparison matrix for features.
Once the technology is determined, organisations work backwards to implement this technology. And determining the intents is very often a guessing game.
Gartner placed emphasis on the fact that organisations must start with defining user intents. And in turn business intents need to align with customer intents.
Cobus Greyling - Medium
Read writing from Cobus Greyling on Medium. NLP/NLU/LLM, Chatbots, Voicebots, Conversational AI, Ubiquitous User…
Get an email whenever Cobus Greyling publishes.
Get an email whenever Cobus Greyling publishes. By signing up, you will create a Medium account if you don't already…
Eliza Language Technology Community - Language Technology: Conversational AI, NLP/NLP, CCAI…
ELIZA - Where language technology enthusiasts unite.
Analysis Of The Gartner Chatbot Deployment Guide
This deployment guide places significant emphasis on Intent Driven Design and Intent Driven Development of a chatbot…
The Customer Service Chatbot Deployment Guide
Over the last 10 years, chatbots have evolved to become a critical technology component of a service organization's…
Will Conversational AI Implementations Become More Fragmented?
Recently I wrote about five emerging Conversational AI trends, the first being Intent Driven Design & Development…
Five Emerging Conversational AI Trends
A recent Gartner report highlighted intent driven design and development as key elements to delivering a successful…