Foundation Conversational AI Technologies Landscape
And the rapid expansion of Large Language Model (LLM) enablement.
The Conversational AI landscape is seeing rapid development on three fronts:
- Voicebot Enablement & Orchestration
- Agent Assist
- Large Language Model Enablement
Version two of the chart below features additional Chatbot Development Frameworks and voicebot orchestration products.
Also, the category of LLM Enabled Chatbot Development Frameworks have been added. This category is sure to grow, and LLM enablement will move from a market differentiator to table stakes.
But…I hasten to say most of these implementations have been very similar, with Cognigy being the exception to some degree. The focus is still very much on the generative aspect of LLMs, and the predictive element is neglected.
While considering and prototyping predictive LLM implementations, I soon realised it is hard to justify the involvement of LLMs in terms of complexity and cost.
As I have argued before, traditional NLU engines yield better intent/classification result in terms of the amount of training data, cost and training time.
💎 One area where LLM implementation leverage and augment existing NLU investment is with a hybrid, NLG enabled NLU…more about this in a following article.
💎 A second area of innovation is enhanced fine-tuning methods for predictive LLM implementations, more about this later.
⭐️ Please follow me on LinkedIn for updates on Conversational AI ⭐️
I’m currently the Chief Evangelist @ HumanFirst. I explore and write about all things at the intersection of AI and language; ranging from LLMs, Chatbots, Voicebots, Development Frameworks, Data-Centric latent spaces and more.
