We are working with a range of companies to develop AI-enabled chat. The audiences and platforms range from B2B commerce to B2C advice, from medical software to travel, from financial advice to shopping. We are working as retained consultant to AI startups and companies providing services a range of services, from strategy to development.
We have broad experience in helping companies understand how they can leverage conversational AI as part of their product and service strategies.
A Bit of Background
Chatbots are not really new (Eliza, one of the very first, was born in 1966 — fairly old). That said, new chat platforms, frameworks, databases, models, technologies and audience technographics add up to a brave new world of AI-enabled chatbots. These range from pretty simple (we’ve all seen chatbots that are anything but intelligent) to super-complicated chatbots that combine loads of data and require advanced technical knowledge to build.
What Are the Components?
Getting an AI-enabled chatbot right is a combination of a lot of things — from the ability to process questions to harnessing necessary data to knowing the audience needs to not letting biases interfere with how a chatbot learns.
An AI chatbot has the following 4 components:
1. Natural language processing and understanding. NLP allows user to talk to chatbots, which in turn review data and talk back. Major advances have been made recently in NLP, as you probably know from talking to Siri or Alexa. Note, machine learning is an important component in enhancing the NLP component.
2. Deep learning is the result of datasets and neural networks that sit behind a chatbot. Learning enables these intelligent agents to see patterns in the information they receive and respond to them appropriately.
3. Knowledge base / knowledge representation covers how the information that is gained and stored. This is all about getting the right datasets and rules behind the chatbot and then programming the platform to learn faster and determine and select relevant information more quickly and more accurately.
4. Decision and action: Chatbots must decide what to do based on all the information gathered and learnt. In other words, like a human conversation, the chatbot met decide what to say next. For the chatbot, the action is just the process of typing out the sentence the chatbot has decided to say.
. . .
“Web Zeit developed a strategy and technology roadmap, with a specific chatbot development and technology plan that provided a cornerstone for our business — and helped us succeed with fundraising.”
– Founder, Travel Startup
“Anthony was a joy to work with. He was able to grasp an understanding of our AI SaaS business within a short time frame, and has helped us refine our product offering and market positioning.”
– Founder, B2B AI Startup
