CASE STUDY Astro enables workplace teams to focus on what's most...

CASE STUDY

Astro enables workplace teams to focus on what's most important by using machine learning to make their communications more effective. Founded in 2015 and based in Palo Alto, California, the company raised $10 million in its Series A venture-capital round. AI simply has to be part of the fabric of any modern app from this point forward. It makes sense to work with existing best-in-class AI services instead of trying to invent them yourself, so it's tremendous that AWS makes them available for us to just plug into." Andy Pflaum Cofounder & Chief Executive Officer, Astro

The Challenge

Email's Not DeadIt's Not Even Sick

Despite frequent predictions of its imminent death, email isn't going anywhere. By the end of 2021, there are expected to be more than 4.1 billion worldwide email users sending and receiving about 320 billion emails a day.

This volume of email has the potential to cause inbox overload, and, therefore, decreased productivity. Current technology solutions include email apps with an increasing number of sophisticated features, and non-email communication apps that aspire to replace email altogether.

Astro, a startup software development company founded in 2015 and headquartered in Palo Alto, California, approaches this problem a bit differently. The company believes that artificial intelligence (AI) and deep-learning technologies like Amazon Lex in the Amazon Web Services (AWS) Cloud can solve the problem of information overload.

Why Amazon Web Services

What Makes Astro Different?

Astro sees immense potential value for email users, hidden right in their inboxes. "Our email accounts hold a massive quantity of untappedand potentially extremely valuable unstructured data and workplace information," says Andy Pflaum, Astro chief executive officer and cofounder.

Astro set out to help email users unlock this value in their email data by using machine learning, the same technology many companies are using to find value in their own unstructured enterprise-wide datasets. "We realized that, with recent advancements in machine learning, we could build an email app that reduces the cognitive overhead of dealing with email," says Pflaum. "By surfacing messages that need immediate attention, learning how people usually deal with those that don't, and getting them out of their inbox as fast as possible, our app helps people focus on more valuable work."

Introducing Astrobot

When Astro released the first version of its Astro email app in early 2017, it shipped with Astrobot, an in-app chatbot assistant. Astrobot uses machine learning and natural-language processing to take on some of that "cognitive overhead" of managing busy inboxes.

Within 30 seconds of first syncing with users' accounts and emails, Astrobot pulls insights and suggested actions such as unsubscribing from lists, automatically archiving certain messages, and highlighting important questions and due dates. Users can also ask Astrobot to set reminders, answer questions, and distinguish VIPs from less important contacts.

Astrobot debuted with a text-only interface, but Astro quickly added an Alexa skill that gave users the option of interacting with the service through voice commands over the Amazon Echo family of devices. Astro decided that the next logical step in the service's evolution would be to offer voice-based email management inside the Astro app itself.

In August 2017, Astro introduced Astrobot Voice, enabled by Amazon Lex, which gives users the ability to manage their email with voice commands without leaving the Astro app. "In many situationsat home, while getting ready for work, or in the car, where you can't safely operate devices by handvoice is the fastest and safest way to retrieve and create information," says Emily Kramer, vice president of marketing for Astro. "We are really excited to be able to offer what we believe is the first in-app voice email assistant, and we couldn't have done it without the AWS Cloud."

Astrobot customers seem to be excited about the new capability, too.

"We saw a 250 percent increase in web sessions on the day we launched Astrobot Voice and acquired three times more new users in the first week, compared to the previous week," says Pflaum, adding that more than a third of Astrobot users tried Astrobot Voice on the very first day it was available.

The Benefits

Deep Learning Enabled by the AWS Cloud

For a startup like Astro to achieve so much, so quickly, it needs IT infrastructure and services that are simple, flexible, and powerful enough to let the company's lean teams focus on code, products, and building value. That's why Astro was born in the AWS Cloud and relies on AWS services for its Astro email app and Astrobot Voice.

At the heart of Astrobot Voice is Amazon Lex, which uses natural-language processing to provide high-quality capabilities for speech recognition and language understanding. With these capabilities, developers can build sophisticated chatbots inside existing applications. The Astro email app also relies on Amazon Aurora to store message and account metadata and uses Amazon Simple Storage Service (Amazon S3) to hold message content and attachments. Astrobot Voice currently uses on-device speech output, but Astro plans to replace that soon with Amazon Polly, an API-driven service that uses advanced deep-learning technologies to synthesize speech and that will convert the solution's output strings into natural-sounding speech interactions with Astro users.

Because Amazon Lex is so simple to use, Astro needed less than two months to add Astrobot Voice to its email applications. "Amazon Lex was the best and fastest way to rapidly deploy Astrobot Voice, especially because we were able to reuse a lot of the same logic we had built for the Astrobot Alexa skill," says Pflaum. "With Amazon Lex, developing and deploying an enterprise-grade voice email assistant took us only six weeks."

That ease of learning and using Amazon Lex and other AWS services also helps Astro achieve big results with its small teams. "By using AWS services like Amazon Lex, Amazon S3, and Amazon Aurora, we only need one full-time employee to manage server operations for an application that is currently available on four mobile platforms and one desktop platform," says Pflaum. "If we had to run something similar in our own data center, managing operations would take at least five full-time employees."

Although some companies are still wondering whether and how AI and deep learning can be useful to them, Pflaum is convinced these technologies are already crucial and will soon be ubiquitous. "AI simply has to be part of the fabric of any modern app from this point forward," he says, comparing deep-learning technologies to earlier basic underpinnings of modern software development, like databases or internet connectivity. "It makes sense to work with existing best-in-class AI services instead of trying to invent them yourself, so it's tremendous that AWS makes them available for us to just plug into."

About Astro

We believe that what makes channels so powerful is a thriving platform that brings together conversations, files, and best-of-breed software tools into one streamlined hub for collaboration.

Benefits of AWS

Built and deployed an enterprise-grade production application feature in only six weeks

Needs only one employee to manage server operations for an application available on five platform

Tripled new user signups the first week Astrobot Voice was available (Article adapted from AWS)

Question 1

With reference to the case study, discuss how the Astrobot assist Astro in customer service. What are the different types infrastructure involved? How do the internet and internet technology work to support communication of Astrobot ?

Question 2

Evaluate Astro principal e-commerce business and revenue models? How has e-commerce affect business to business and consumer to consumer transactions?

Question 3

Analyze business intelligence environment of Astro? How do business intelligence and business analytics support decision marking? Should Astro use business intelligence for customer service and marketing? Why or why not? What kinds of companies are best suited to use these systems?

(It is essay Answers Question)

(reference source needed)

Astrobot Voice Architecture Alexa Amazon ElastiCache HTTP Session Data GRPC PostText STT TTS Intent Astro Android Elastic Load Astrobot Voice Amazon Lex and iOS Apps Balancing System Astro Email Gmail or Office 365 Service Amazon Aurora Amazon $3 bucket