“Mr. Watson–come here–I want to see you.”

One of the fun parts about being an entrepreneur is venturing out into new territory; to (try to) keep up with the breakneck pace of technology and see where the intersection lies between what new tech you can play with and how it overlaps with what you actually need to do.

Recently we started playing with some new Artificial Intelligence “AI” technology from IBM Watson, and we learned a few things.

1. Image recognition is really fun. Funny actually.
2. Its incredibly easy to try out.
3. We have much more to learn.

Approaching AI has always been a little daunting. When I was growing up, the term “AI” always came hand in hand with “humanoid robots” and “singularity”, which required science well above my education level or comprehension. The good news is that companies like IBM, Google, Microsoft and Amazon have broken down the topic into smaller disciplines, making AI much more approachable and usable, even though I still have no idea what’s happening under the hood.

IBM Watson, for example, has 19 distinct “AI” services to choose from. Microsoft Cognitive Services has 30, everything from “Emotion APIs” – which determine the emotion of the content you pass to it, to a “Custom Decision Service” which uses “reinforcement learning in a new approach for personalizing content”. You could spend hours exploring all the different services.

For our first foray into AI, we decided to do a quick integration with Watson’s image recognition service. This is a service that you send images to, and Watson will process the image and return a few scored keywords based on confidence intervals. Here’s one of our first tests:

Pretty cool, no? We sent a picture of a mug to Watson and Watson told us that it was pretty sure it was a coffee cup, maybe a mug.

So, then we pushed Watson a little more, and here’s what Watson thought about a picture of Pete, our CEO:

Not as confident this time… Watson looks to be somewhat sure that the picture is of a physical therapist, or maybe a middle aged scholar. While it’s not perfect, it’s not totally inaccurate. Let’s just say we’re excited to see how the technology progresses.

How we did it

What made this project especially exciting was how easy it was to put together.

First, using a simple form and our newly released ‘flows’, we mapped out an experience where users would tap a button to pull up a simple photo form on their phone. Then, using the camera, they could take a picture and click ‘Done’. We used a simple integration script to pass the photo from TrackVia to IBM (about 100 lines of javascript) and then TV saved the results and presented Watson’s findings to the end user. All of this would happen in seconds.

All in, we were able to throw this project together in just a few hours. Even so, what can you do with this information?

First and foremost, with image recognition, you can make your images searchable. For that matter, you can make your audio and video searchable as well. Not only that but you can auto-translate these images, audio and video into different languages.

The short answer is, there is a ton you can do with this new technology. And we can help. Would you like to try something out? We’d like to help. Take a look at the services below. Let’s make something epic!

Microsoft Cognitive Services
https://www.microsoft.com/cognitive-services

IBM Watson
https://www.ibm.com/watson/developercloud/services-catalog.html

Google Cloud Platform
https://cloud.google.com/products/

Amazon AI
https://aws.amazon.com/amazon-ai/

Subscribe to TrackVia’s Blog