Transforming the Employee Experience with Alexa for Business

voice and chat

At Kopius*, we help enterprise customers worldwide understand and apply next-generation technologies in smart and innovative ways to advance their business goals. These goals often include improving the employee experience.

Across modern technologies such as voice and chat, artificial intelligence, robotics, augmented and virtual reality, blockchain and more, Kopius has a wide range of innovation and consulting capabilities that deliver forward-thinking, quick-to-market and cost-effective solutions to our customers.

One of the mantras we follow is “learn by doing.” We believe that by providing our own employees with instantiations of the same solutions we build for our customers, employees will not only understand our business better, but will be more interested and involved in what we do.

This is why we built the “Valence Bot” for our employees — to provide them with a tool that not only helps them with their daily jobs, but also helps them understand modern voice and chat platforms — one of our key technology pillars.

What is Valence Bot?

Valence Bot is a chatbot that uses Amazon Alexa, Amazon’s virtual assistant, to answer employee questions and handle requests related to the company. Employees can obtain company information and make requests simply by interacting with Alexa from any personal Alexa device. This includes things like:

· Obtaining employee benefits information such as healthcare provider information, group policy numbers, and paid time off (PTO) information.

· Requesting IT support or submitting facility requests.

· Accessing personalized information based on job role, such as company sales data for the Executive team, or new hire data for the Recruiting team.

How Does it Work?

Valence Bot is built on top of the core framework outlined in the Question and Answer Bot blog entry originally posted a year ago in the AWS Machine Learning Blog. After adding additional functionality and integration points to this framework, the solution is bundled into an Alexa private skill and deployed to employees using Alexa for Business. Every new Kopius employee is given an Amazon Echo device during onboarding. They are also enrolled into our Alexa for Business organization, allowing them access to the private Valence Bot skill from anywhere.

The high-level architecture of Valence Bot is shown in the following diagram:

Details of each component are as follows:

  • Alexa Device — A personal Alexa device that employees use to interact with Valence Bot. Kopius gives all employees an Echo Dot, but they can access Valence Bot through any existing Alexa device they may already own. Access through the Alexa mobile app is also possible.
  • Valence Bot — This is the private Alexa skill that gets deployed to each employee that serves as the user-facing interface to information and requests.
  • Alexa for Business — The management layer that is used to deploy the private skill to all employees.
  • Chat Engine — Built on Amazon Lex, the chat engine is ultimately what gets exported to the private Alexa skill. Lex is configured with just one slot and intent to capture the text of the question for further processing.
  • API Services — A web interface (protected by Cognito authentication) accessing an Amazon API Gateway instance provides administrative access to manage content served by Valence Bot.
  • Content Datastore — An Amazon Elasticsearch instance provides the mechanism to search the question and answer data for the best responses. In addition to simply returning an answer, a “hook” can be assigned to questions which enables custom Lambda functions to be triggered for the purpose of obtaining additional information or handling requests.
  • Fulfillment — AWS Lambda functions are used to serve two purposes in this architecture:
    • Interface from which Valence Bot interacts directly to provide answers to questions and fulfill requests. In this scenario, a Lambda function queries Elasticsearch and returns the best answer.
    • Seeks out additional information beyond simple question and answer data in Elasticsearch. In this scenario, a hook defined in Elasticsearch triggers additional Lambda functions that then fetch additional information or handle requests. Example include integration with 3rd party APIs to obtain sales data or submit service requests.
  • Notifications — Amazon notification services such as Amazon SES and SNS notify the user when appropriate, such as when their service request has been submitted, including information around expected SLAs.

What Do Our Employees Think?

Teah Delfino, our Director of Recruiting, explains the benefits of Valence Bot from her perspective:

“When I brought it home and set it up, I thought I wouldn’t use it much — I’m in HR so I have access to a lot of company information. But I find I use Valence Bot all the time — it’s so convenient to simply say, “Alexa, ask Valence Bot when our health insurance renews,” or “Alexa, ask Valence Bot who we’ve hired in the past month.” It’s also a compelling story to tell when I’m speaking with potential employees. I get a lot of positive feedback about what a unique benefit it is and how excited they are that we live in the world we work.”

Interested in Learning More?

Interested in learning more about Valence Bot or how voice and chat technology could improve your employee’s experiences? Contact us and we’ll start you off with a demo, to show how remarkable this technology can be!

*Kopius performed this work under its previously known business name, Valence.

Additional Resources: