Getting started with Watson IoT

By | 4 minute read | June 3, 2016

Much has been written about how the Internet of Things has disrupted traditional businesses — and how that disruption is poised to skyrocket as we distill better insights from our connected devices. Today, smart leaders already know it’s time to get started exploring the power of IoT for business. What most don’t know, however, is how to actually get started.

Last month, I gave a talk on this very subject at Reliability 2.0 — a gathering of thought leaders in the manufacturing and asset management space. There was a lot of curiosity about what Watson is and what potential it has to serve maintenance and reliability applications. Many of the comments I received after the talk showed that attendees were excited to hear how easy it was to get started. Everyone knows how to save a list of data in a spreadsheet format, but few know that the Excel file could be uploaded to Watson Analytics to start exploring the data for information. They were happily surprised to learn they could dive in and experience Watson IoT for themselves right now (no Computer Science degree necessary).

Here are three easy steps manufacturers can take today to kick off their IoT transformation:

Step 1: Connect and talk to Watson

Embarking on an IoT strategy can seem daunting, but its easy to start small. Watson provides user-friendly, advanced, collaborative tools to analyze exported Maximo .csv files. As a cloud-based offering, no complex installations or integrations are required. Several analytic models are already there. Upload some of your maintenance history data from Maximo and start exploring. It’s as easy as signing in to the Watson IoT platform and uploading a .csv file. Watson can even tell you if the data you’re collecting is able to be analyzed; data health is one of the models available on the free site.

Step 2: Mobile notifications

Compress the distance between maintenance and the health of the assets they are responsible for by sending mobile notification to your workers.

If you’re like most businesses, up to 50% of your maintenance efforts have zero impact on downtime (ARC EAM Market Research Report 2014).

The promise of Cognitive IoT is that it can help manufacturers work smarter — not just harder — by using data to optimize your preventative maintenance.

Send a mobile notification to your worker when a device is close to failing. For example, if your sensor is monitoring a spindle or a pump, this might be when it wobbles or vibrates. You can set these rules and fire off the notification easily from within the platform. Closely track how much downtime you’re eliminating by improving your preventative maintenance. Even a few minutes shaved off small production delays add up quickly. Recapturing a few minutes per day on a single line can translate into significant uptime improvements. Small decreases might not seem groundbreaking on their own, but they quickly add up.

Step 3: Let Watson work its analytic magic

Bring your data and see what Watson can uncover. When you first signed in to Watson and loaded your .csv file, you saw the results of analytic modeling across all the rows in your spreadsheet. You are then invited to ask questions – Yes, ASK questions, using normal human language. Questions like: is there a correlation between the time of day and the mean time to repair on this asset? Or many assets? Who is most effective at planning a pump replacement? How often is a failure in this area of my asset actually caused by some other connected asset?

The real power of the Watson platform isn’t in finding what you’re looking for, but revealing what you didn’t know was effecting your business. Maybe it will find that other devices are showing similar signs of failure. Maybe it will find other related, problematic patterns based on the data you provided — rocks you would never think to look under.

Now, take a step back and look at what you’ve accomplished. It’s possible that some of the early explorations of Watson have given you ideas on where you might apply it. My best advice is to find a high value project that is small and focused to apply the first test of IoT, Maximo and Watson Analytics. Don’t try to boil the ocean or take on too big of a scope. Take the time to learn on the small efforts, capture and socialize the results. Choose your next target carefully.

By closely tracking the downtime you’ve eliminated over the course of these experiments, you’ve essentially laid out a business case for further expansion of IoT. And, by exploring Watson views of your Maximo data, you’ll have brought a cutting-edge, cognitive computing element to your work – investigating patterns and insights you may have never discovered otherwise.

Want a deeper dive into getting started with Watson IoT? Watch a full video of my talk below.