After completing the AWS AI Practitioner course, I wanted to test my knowledge of the services, and create a chatbot that would be used via Slack.
These are the requirements I went into the project with:
- The bot needed created using AWS Services
- The bot needed to utilise GenAI for interactive discussions
- The bot needed to link in with Slack
- As well as GenAI, the bot needed to have access to domain based data
- The bot must be easy to use Bonus - If I could get the bot to work via Alexa
After reviewing the options, the following services required greater analysis due to GenAI/ChatBot features, and data indexing:
Amazon Kenda - Kendra at a high level provides indexing via Amazon S3 or via API with large etxternal companies such as Google, Microsoft and Atlassian.
Amazon Lex - Lex is a chatbot that can be configured with custom intents, allowing for a personal end user experience, with the ability to bounce user queries.
Amazon Q - Is a managed subscription based chatbot, allowing for users to interactive with the client within web or mobile UIs.
The next phase will involve a deep dive exploration of the key features, customerisation and cost of each.
📝 Note: When exploring these options, it was important to weigh up cost v feature sets 📝
Amazon Lex: Build bots with Conversational AI
Amazon Q: The most capable generative AI–powered assistant for accelerating software development and leveraging companies' internal data
Amazon Kendra: Find answers faster with intelligent enterprise search powered by machine learning