A little background
Around the start of Q2, major technology companies host "summits" where they announce new product offerings, have breakout sessions and throw crazy parties. The events are covered in great depth by the media and many attendees analyze company strategies for the year ahead based on these events.
Frugality as an advantage
Since Cloudy has no budget, I can’t travel to any of these - which is an advantage. Instead, you can watch replays of the event, from the comfort of your own own - at 2X speed. If you’ve never watched a YouTube video or listened to a podcast at 2X speed, well, it’s interesting. At times your brain wants to explode and life feels much slower after you stop. But, the advantage, is you can watch twice as much content then if you attended in person.
After watching numerous company keynotes and breakout sessions, a commitment to AI/ML is certainly a theme to all of them. Google announced 29 new products related to AI/ML. Adobe dedicated a significant chunk of its keynote to “Sensei”, their own AI. Ginni Rometty’s keynote address was called “Building Cognitive Enterprises.” Most of the presentations have the same feel and many trends, which we review below, are starting to emerge.
1. AI for Everyone
If you don’t believe me, Google had a slide in its keynote that said “AI is for everyone.” Companies are starting to build products that allow everyone, regardless of programming experience, to use AI/ML For example, Google now has a point/click feature where you just select from a dropdown what you want to predict, which columns you want to use, and voila, you have a machine learning model
Example of Google AutoML
Adobe is doing the same thing. Shown below (sorry for unreadable image, their color choice, not mine), the automatic ML offering works by predicting things like churn and upsell opportunities for different people in your CRM. Adobe takes it one step further and gives you a column of what variable is the most important for that prediction. For example, if the ML algorithm scored a user as “high risk to churn” it also tells you why the algorithm made that prediction. Possible examples are “Low Product Usage” or “Geo”. The idea is, if a consumer had “Low Produce Usage”, one could personalize a campaign that mentions different ways to use the product.
2. AI Everywhere
When you think of AI at home, Alexa and Google come to mind. When you think AI for business, Microsoft wants you to think about them. It makes sense; almost every large organization uses Office 365 and having all of that data puts Microsoft in a good position to develop AI around how we work. For example, Microsoft showed a video of Cortana, their virtual assistant, who will help you manage your schedule, book conference rooms and invite people to meetings. Microsoft is also putting AI into programs like Word, which is adding a feature that checks how clear your writing is and suggests new ways to phrase sentences.
Watching the Microsoft Build keynote, I got the feeling Microsoft is trying to be an IBM Watson competitor. A lot of the case studies that Microsoft reviewed felt quite similar to cases that IBM Watson has presented. While Microsoft doesn’t have a “Watson” product, they are helping a variety of companies across many industries integrate AI.
3. Vertical Focused
Retail was a big area where companies were applying AI/ML solutions. A great example is Google’s new product, AI for the call center. If someone calls a Google enabled call center, first, Google has a bot try to answer a caller’s original question. If that doesn’t work, the call is routed to a human, where the human is given suggestions by AI/ML about how to solve the problem. The AI/ML continues to listen to the call and if the customers problem changes, the computer gives new suggestions to the human in real time. Call center agents also get reports on what responses worked well with different issues, which is also based on, you guessed it, AI/ML.
Adobe highlighted how Foot Locker is using AI/ML to improve the in-store experience. They cited an example where users could hold their phone up to a shoe wall and using AI/ML, the phone would identify what the shoe is and if the current location has it in stock. Instead of waiting for a store assistant to help you, voila, you know right away
4. RPA gets smarter
AI/ML is starting to be integrated into robotic process automation . IBM announced “IBM Business Automation Intelligence with Watson”, which involves a whole range of solutions from identifying where to automate, to implementing RPA, to measuring its impact. IBM gave an interesting example where RPA is predicting if invoices are fraudulent instead of just processing them. It seems like machines are starting to make more and more decisions.
UIPath also announced a feature which allows RPA to operate via virtual desktop environments. The issue with virtual environments is RPA cannot process it because it is an image of a screen. UIPath built a neural network to navigate within the video screen of the desktop and a blog post by UIPath called it a “true breakthrough for the RPA industry.”
The Final Word
AI is anywhere and everywhere and now becoming easy to use. If you haven’t used these tools, start small, pick a pilot project to see what is possible.