2020 has seen accelerated growth in AI, specifically the Virtual Assistant space.
Last year, I transitioned from a business analyst to an AI Engineer developing Virtual Assistants.
Here is what you need to do to get up to speed and incorporate into your development plan before year end!
- Understand core terms and architecture
A Virtual Assistant (VA) is a Virtual Agent that provides information (and perhaps, automation) on a topic.
Users interact with Virtual Assistants using utterances (phrases). The utterance should trigger an intent, the topic the user is looking for. Within the intent, the user may be asked to provide additional information (slots) required to fulfil the intent and process the request. Containment metrics, the rate at which the bot can accurately link an utterance to an intent, is an important metric used to track bot performance.
I recommend reading the Age of Intent by P.V. Kannan to get an overview of Virtual Assistants and their capabilities.
Having worked on two Virtual Assistant platforms, understanding these core terms is fundamental to understanding how this technology works – giving you the ability to socialize AI within your organization and ultimately, shift your industry.
- Implement based on an (agile) strategy.
Start with a specific user case and build incrementally.
Build a proof-of-concept (POC) based on your use case.
In AWS Lex, the technical architecture will probably look like this:
An AWS Lex bot link to an AWS Lambda which will fulfil the intent. The AWS Lambda will retrieve information from a database or other AWS database service.
Next, jump into an AWS tutorial and build your first bot! Then, check out more sophisticated examples on the AWS GitHub for other POC ideas.
3. Share key learnings.
This could be in a lunch and learn, technical knowledge share or even part of a retro.
Share your use case, initial tech architecture and any customer input received.
This is a great opportunity to develop your technical storytelling skills, create compelling visuals that explain you POC and receive any feedback from stakeholders.
AI will continue to grow in 2020 and beyond. By incorporating these ideas into your development plan you will be able to explain key VA terms, demo a working POC and share your product journey with others as a thought leader.