Key Features of Conversational AI Platforms
By Alessandro Botticelli -- November 05, 2019
We build a lot of different types of chatbots at The Bot Forge and deliver these to a variety of channels such as websites, Facebook Messenger, Slack, and WhatsApp. To create our chatbots we often use different AI platforms which offer more suitable features for a specific project.
Major cloud and open-source providers have implemented comparable feature sets for conversational AI platforms with strong natural language understanding capabilities, along with solid open-source privately hosted alternatives.
API and UI
A conversational AI platform should provide user interface tools to plan conversational flow and help train and update the system.
Context
Beyond intent and entities, context objects enable systems to track conversation context, user situation information, and conversation position -- vital for complex conversations beyond simple FAQ bots.
Conversation Flow
The system considers conversation position, context, and user utterances with intents and entities as rules managing flow. Platforms offer flow engines and visual tools. Slot-filling ensures entities are present and prompts for missing ones.
Pre-Built Channel Integrations
Platforms supporting target channels accelerate chatbot delivery and enable channel flexibility for the same conversational engine. Dialogflow is highlighted for strong tooling in this area.
Chatbot Content Types
While conversational AI focuses on text, messaging systems involve buttons, images, emojis, URLs, and voice input/output. Platform support for these features creates richer experiences and manages conversational flow.
Integrations
Bot responses improve through integrating user information with internal or external web services. This is one of the most powerful chatbot solution features.
Pre-Trained Intents and Entities
Some systems provide pre-trained entity types like dates, places, or currencies and common user intents like small-talk rather than requiring project-specific creation.
Analytics and Logs
Successful chatbots need constant training and monitoring. Platforms should provide dashboards showing user conversations, response statistics, user interactions, and metrics. Log exports aid system improvement by tracking missed intents, negative sentiments, and flow issues.
Tech Stack
Library availability and supported languages matter. Node.js is our server stack choice, with most AI platforms supporting it.
Costs
Cloud hosting and NLU solution costs are critical, particularly for enterprise chatbots handling large traffic volumes where monthly NLU costs reach thousands of pounds. Free tiers typically exist with paid tiers offering enhanced services, greater volume support, and performance.
The Platforms
Platforms planned for deeper analysis include: Botkit, Chatfuel, Amazon Lex, Microsoft LUIS, Google Dialogflow, Rasa, and IBM Watson.
Want to build with conversational AI?
Book a short call and we will help you choose the right approach.