Blog: Developing a Smart Meeting Logger with Amazon Alexa Cloud Service

The Corporate World runs on meetings. Some meetings are vital to the running of a Fortune 500 company and others could be replace with simply an email. However, until the day comes that we can link minds together and have all meetings telepathically, they’re going to be a staple of modern working life. Meetings themselves can be a logistical nightmare. Trying to organize where, when, and how everyone meets can make preparing for the meeting as difficult as the meeting itself. While the advent of Skype, Chime, Slack, Teams, and Hangouts have made it easier for people to join meetings from wherever they are in the world, most meetings are still held in small, dimly-lit conference rooms across the nation. To help fix this problem, the DB Best IoT Hackathon challenged our teams with finding the best way to make corporate life easier using Amazon’s Artificial Intelligence, Alexa.

Smart Meeting Logger

Meet the winner of DB Best IoT Hackathon — the Smart Meeting Logger! Our solution utilizes the Amazon Alexa cloud service and allows you to seamlessly use an Amazon Echo to log the meeting.

To activate the Smart Meeting Logger, all you have to do is tell Alexa to turn it on. From there, everything works with your voice. Registering meeting participants and attendees, creating the meeting agenda, and itemizing topics of discussion are all done effortlessly by talking to Alexa. During the course of your meeting, each item that you discuss will be cataloged by the Smart Meeting Logger and emailed to the participants at the end. Future iterations of the skill will be able to record the entire meeting so it will be like having your own private stenographer in the room with you.

At the end of the meeting, the Smart Meeting Logger uses Amazon SES service to send email notifications to all meeting participants.

These are the primary capabilities of the meeting logger service, developed by our hackathon winners. The team did a wonderful job, beating other great solutions. Here’s a brief description of other Alexa skills, created by the other DB Best Hackathon teams.

  • Skill to search for developers with certain skills in our internal personnel management system
  • A skill for our developers, connecting Alexa to GitHub. It allows you to read your personal to-do list, close to-dos by position, as well as reading pipelines on user projects and merge requests
  • And, to round the things up, a cool email checker

Now let’s talk about creating the Amazon Alexa Skills in general.

Background on Amazon Alexa Skills

Basically, Alexa is Amazon’s cloud service with smart capabilities. Alexa is the brain behind various smart devices including the Echo family.

Amazon Alexa supports voice interaction and allows for playing music, setting alarms, as well as providing weather, traffic, and other real-time information. Also, Alexa can control other smart devices in your home automation system. However, in addition to these built-in Alexa skills, you can create your own custom services. You can build new skills using the Alexa Skills Kit (ASK), a collection of self-service APIs, tools, documentation, and code samples.

Actually, our hackathon teams had to develop a new Amazon Alexa skill. Now we will cover the winning solution in details.

Architectural diagram

First of all, let’s take a look at the architecture diagram of the Amazon Alexa skill we created.

Architecture diagram of the IoT solution with Amazon Alexa

First, we place the Amazon Echo Dot in the meeting room. The Echo Dot interacts with the Amazon Alexa cloud service. Alexa recognizes the skill you want to use and calls the relevant Lambda function. This function interacts with Alexa, S3, and SES. As you can see, all services live in the Amazon cloud, so our solution is both very flexible and highly reliable.

In addition to these benefits, we want to highlight another advantage of cloud architecture. You can replace any of the services we utilized by simply changing one or two lines in your code. For example, we can change Amazon Simple Email Service for a more advanced email sender literally in a couple of minutes.

The Meeting Logger skill supports sessions, and this is one of the defining features of our solution. This means that you can restore the session even if your Echo Dot goes down. So, in the event of a power outage, none of your meeting logs are lost.

Step 1 — Names recognition in Amazon Alexa

Amazon’s smart assistant does an amazing job with voice recognition. However, this is the case often for the short commands, but not for long and complicated surnames. We needed to add the meeting attendees with voice commands, so, we faced the most difficult challenge with the correct surname recognition.

Alexa recognizes the voice commands in the cloud. However, users can fine-tune this process. For example, you can set the string similarity level to diminish the influence of pronunciation issues. Also, you can specify the key part of the word for smoother names recognition.

We spent quite a lot of time tuning this voice recognition service, and the results proved to be amazing. During the presentation session, everyone could talk to Alexa and the service worked perfectly.

Step 2 — Amazon SES Email verification

Amazon Simple Email Service (Amazon SES) is a relatively inexpensive cloud-based email sending service. However, Amazon SES has a strict requirement that makes the developer’s life a little bit more complicated. Actually, you need to verify all emails before you can use them in your cloud service.

This may not look like an issue for massive production services running in the cloud for years. For this quick and dirty hackathon solution, we decided to stick with a limited number of participants with previously verified emails.

Step 3 — Amazon Alexa Development process

You can create new Alexa skills through the Alexa Developer Console. Here you can also debug your Alexa skills. The picture below shows how the Alexa Developer Console looks inside (click the image to enlarge).

Alexa Developer Console

You can also utilize Alexa for Business to develop an enterprise-level service with Amazon Alexa. However, our hackathon teams considered it rather expensive. Also, they noted that Alexa for Business solutions require a specific application architecture. So, this path would not lead to the flexible solution that our teams were looking for.

Alexa lives in the cloud and takes care of all voice recognition activities. This means that Amazon Echo Dot and other devices compatible with Alexa simply record your voice commands and send them to the cloud.

Usually, you may want to use AWS Lambda to build and host your new skills. The Lambda function you create then interacts with other Amazon cloud services. Particularly, in our case, the Lambda function decided whether to provide a response to the user, store the data in the Amazon S3 bucket, or call the Amazon SES to send our the email notifications.

We implemented 15 event handlers in our Alexa skill. This means that our solution supports 15 different voice commands.

At the end of the hackathon, our team delivered a fully working prototype of Amazon Alexa Smart Meeting Logger skill. You can find the source code of this amazing cloud-powered meeting logger on GitLab.

Future steps

Very often the developers create raw solutions during the hackathon. Our Smart Meeting Logger is not an exception. So, we plan to upgrade the solution we developed. Particularly, we want to add the following features.

  1. Admin panel that allows for adding pre-defined groups of users (for example, we can add all iOS developers to the meeting with just 1 voice command)
  2. This admin panel should also support creating and editing the meeting agenda before the meeting starts. So, you don’t need to set it up at the start of the meeting
  3. Add the recording capabilities. This will increase the maintenance cost for cloud storage. But a full-service smart meeting recorder is worth it!

In addition to these great ideas, we can use Alexa to manage other smart devices in the meeting room. For example, Alexa can adjust the lights in the room and turn on the projector in case you need to share a presentation. Also, we can set up the camera for face recognition. In this case, the Smart Meeting Logger will recognize the meeting attendees and add them to the list of participants automatically.

How we created the Smart Meeting Logger

While the teams were heads down with coding, the mentors had some time to share their thoughts. Check out what the mentor of the winning team told us.

The idea to create a Smart Meeting Logger was not long in coming. Every day we attend meetings, and every day we need to write up and analyze their summary.

This is a rather long and complicated process, which may even not happen at all. The meeting logger that we created allows for faster creation of the meeting notes and makes this process very convenient and just fun.

Our team developed a fully working prototype of the Meeting Logger service. However, we can expand its capabilities in the future. For example, we can integrate it with the DB Best’s internal personnel management system to implement the search system. Also, we plan to add an admin panel to pre-define meeting participants and agenda. And this is just the tip of the iceberg.

Every member of our team dreamed of creating a new skill for a voice assistant, and the DB Best IoT hackathon provided us with such a chance. We figured out that creating new skills for Amazon Alexa is not that hard, it’s rather more fun process.

Also, I’m very happy that our team members had a chance to prove themselves in extreme conditions. I think all of them were able to open new frontiers with this event.

Iliya Sokolov, the mentor of the winning team, .NET developer

Thanks for changing office life for the better

DB Best IoT Hackathon brought our top-notch developers together to in a desire to develop a truly unique IoT service. The organizers provided all participating teams with an Amazon Echo Dot. This Bluetooth-enabled smart speaker supports Amazon Alexa virtual assistant. The teams had to create a useful service for better office productivity within a limited period of time.

Wee also want to thank all teams for their desire, commitment, and willingness to change office life for the better. We highly appreciate your desire to create state-of-art IoT solutions as a part of DB Best IoT Hackathon.

Contact DB Best to start developing a new Amazon Alexa skill or even a custom smart device from scratch! In the DB Best R&D Lab, we have all the necessary equipment to bring your most crazy ideas to life.