March 18, 2016

The Rise of Software Bots — The Applications That Take Big-Data to the Next Level

—— written by Matt Cheng, Founder of Cherubic Ventures



I think software bots, powered by big data, machine learning, and AI, will be the new communication trend since websites and apps.

In Bots World, No Interface is the Best Interface

Take “Penny” as an example. According to the iPhone app description, Penny is a personal finance app and “money coach” (yes, when you build stuff for the millennials, you must dumb down your word choice.) Just like Mint, Penny is a personal finance tool that is connected to your bank account and record your spending. The key difference is that Penny is completely integrated into a chat interface. Users don’t need to navigate in the complicated spending summary or charts. All they need to do is chat with Penny, the bot, which just feel like a real bank assistant. Users can tell Penny what they want to know by clicking on a couple predefined commands, such as “my spending graph”. Penny will even tell you things you didn’t know to ask such as: “on average, you eat out five times a week!”.

Image for post
Resource | Penny’s chat interface

One of the secret ingredients that makes Penny so easy to use is that it limits users’ inputs and provides a guided conversation rather than allowing users to ask random questions. In the past, many companies’ approach to building a virtual assistant’s is to try to collect and analyze all the data and hope it can answer anything a user asks. However, it turns out that users are clueless about what they want or simply do not know how to use the data. This is a main reason why so many AI or virtual-assistant products have failed. Another example: Apple’s Siri. When Siri was first launched, it received lots of hype. But no one really relies on Siri to manage their daily routines today. So I think the key to such products is to provide guided conversation to deliver better user experience.

Using a fitness tracker to track your exercise and calories is nothing new.

“But imagine if a bot can tell you that you’ve only exercised 3 hours this week, which is 5 hours below your exercise goal; and then the bot recommends nearby gyms to you based on available time shown on your calendar.”

Wouldn’t that be much more useful? This is what users really need.

News company Quartz has also jumped on the bandwagon and launched a news app that basically is a bot messenger. The app literally texts you news, along with GIFs and emoji, and in the process figuring out a user’s news preference and then deliver personalized news stories to them.


The Ecosystem of Bots

Standalone bots such as Penny and Quartz app are great. But another type of bots are also gaining momentum: the messenger bots. There are many third-party bots that exist solely on large chat platforms to take advantage of platform’s massive user base. Even platforms themselves such as Facebook, Slack, WeChat are also busy building their own bots. Telegram already has thousands of third-party bots that act like widgets, performing any tasks ranging from searching funny GIFs to checking weather. Bots on Slack are much more business-oriented: such as Nikabot, which monitors team performance; birdly, which automatically aggregate travel expenses after scanning receipts; Statsbot, which analyzes data fed from Google Analytics. Bots on WeChat are usually half bots, half human, providing on-demand customer service such as booking flights, ordering food etc. Facebook by far is the most ambitious. At its developer conference F8 last year, it announced the plan to open up its Messenger appstore and started integrating with many third-party apps. Many of them are content tool, such as GIPHY, which is one of my favorite. Business tools are also a big focus for Facebook Messenger third-party services, since many businesses use Facebook as the primary tool to communicate with and acquire consumers. I think we will very soon see many “business bots” performing tasks such as help you shop or track your orders in the Facebook Messenger app.

Image for post
Resource | Bridly interface

Given the chat interface, the first wave of Bot applications are mostly built upon messenger apps. However, as more programmers delve into making bots and more open-source tools become available, I believe we will see more diverse Bot applications for a wide range of verticals, especially in the fields of IoT and IoV.

Another key reason that makes Bots so popular is that, for millennium or younger generations, most of their social interactions happen online or virtually. Once VR becomes mainstream, we will see even more activities taken place in the virtual world. Software needs to adapt to this new trend, having an “interaction” interface will be the key in winning customers.

“IoT” and smart devices have been the darling of the tech industry in the past few years. Every smart device has its own app that promises to track every details of users’ lives. But as people ponder how to integrate all these apps, perhaps the most important question is how to leverage these massive data in a way that’s meaningful and actionable to consumers. I believe bots will soon emerge as the new trend to provide data and services to consumers.

Recent Articles

See All