Artificial Intelligence and Machine Learning in Next Generation Workspaces

Welcome back to part two of this eight-part series! Now in part one, I introduced the idea of a digital transformation phenomenon that is happening in our world today. I talked about how it is affecting us as consumers and in the enterprise as well. I briefly touched on some of the drivers that are pushing us toward this digital transformation. Those drivers are bringing about some very important and disruptive trends in the realm of digital enterprise. Below is a snapshot of these trends.

Today I will discuss the first of these trends, Artificial Intelligence (AI) and Machine Learning. AI seems to be all around us already as we see today in countries like the U.S., there are millions of consumer devices with embedded intelligent assistants. Like smart appliances for the kitchen or assistants like the google home or amazon echo.

With this digital transformation, more and more intelligent things will emerge in industries like healthcare. For example, MRI image analysis for unique types of cancer which is difficult for human doctors to identify. There will be better analysis via these machines. We will see more and more AI use in industries like healthcare or manufacturing and greater investment in those areas over the next few years.

As we move forward AI will be applied in a focused manner. It’s not about using these technologies for general purposes – we are far away from that. Implementations will be for well-scoped purposes and with clear goals in mind.

Another aspect of learning to adapt to AI is dealing with complexity and having systems that learn and predict or adapt and act in ways that were not explicitly programmed. More implicit programming model with feedback loops would be necessary so the system can act autonomously.

A critical point to remember is that these are model driven systems. Typically, we explicitly program software or a machine to act based on rules we give it. In other words we tell them what to do. But in these advanced systems, instead of an explicit rules-based method, you build a model that understands some purpose and you feed the system data (where the content acts as the code) and the system learns from that data and eventually operate with little or no human input or guidance. An interesting idea to comprehend is that if a system makes a mistake, it’s not actually a bug, the system just hasn’t learned it yet.

This is what makes the smart machines appear “intelligent”. Instead of explicitly defining the rules, your defining how it’s going to interpret the data. The big questions enterprises need to ask themselves is: do they have people to understand that concept? Have they thought about how to go about training these systems?

Over the next few years, I think we’ll see a number of enterprises employ dedicated people to train neural networks. This notion of not programming the system, but training the systems will become an important concept. We’re going to build deep neural networks and drive this to a more context and event-driven model.

 Intelligent Things and Applications

AI & Machine Learning will drive the development of intelligent things and apps. With respect to intelligent things one implementation is robotics into various workspaces in the future.

In retail, for example, Lowes Innovation has a hardware store application – it’s a pilot in about 19 stores. The app has customer engagement capabilities like speech recognition. When you ask it “I need some nails”, it understands your request and with its product info/location capability it can take you to or direct you to the appropriate aisle in the store. It can also scan and check inventory in the shelves to see if any items are out of stock and initiate a restocking procedure.

Hospitality Robots for hotels and cruise ships and other workspaces are also emerging. For example, in a hotel in Nagasaki, Japan, hotel guests are greeted by 10 lifelike robots, removing all human staff from reservations. Butler services robots are also in place at some hotels. Robotics in healthcare will deliver medications and supplies to doctors and nurses inside the hospital.

Now these are specific use cases and implementations but they will drive robotics forward to permeate in more general scenarios. We must start thinking about how these innovations will impact our workspaces for tomorrow – whether in the warehouse or hospital or in the carpeted office.

Another example of intelligent things, with the notion of leveraging artificial intelligence, is autonomous vehicles for specialised environments. Remember workspaces for tomorrow are not all within an office building. Mature implementations exist within farming, mining and warehousing.

Companies are saving hundreds of millions of dollars by using autonomous trucks, autonomous drilling mechanisms for mining of ore and autonomous trains getting it to the ship. Further down the road you could see autonomous ships! (Though that would be longer term given technical issues with bandwidth and security issues.)

Farm tractors operate in a coordinated fashion on farms. These specialised and controlled workspaces are the low hanging fruit for these types of intelligent things because they are being used for a very specific purpose. In industrial settings, vehicles can more fully autonomous. But in general use as of today, we are working with the assisted model pictured below on the left. It’s not true autopilot…yet!

In the future, there’s the vision of fully autonomous. But know that the issue will not be technology holding it back, it will be regulations, liability, insurance, etc. In the meantime, government is trying to accelerate this process and we will progress to the image on the right above where there is no steering wheel and the passengers are not even looking at the road!

Organisations need to be thinking: what does this means for us even if we are not in the auto industry? What if this becomes the rolling office where my employees can work? Or the rolling endpoint? Or the virtual environment where I can deliver consumer products or information? How will this directly or indirectly impact our business?

Organisations are applying AI and machine learning to create new application categories as well. Intelligent apps have the potential to transform the nature of work and the structure of the workplace. Some types of intelligent app focus specifically on operational efficiency. Here’s an example from McDonalds. This is actually from 5 years ago and is driving greater efficiencies over human or manual processes.

The opportunity or challenge here was to move from manual to automated inspection of burger bun production to ensure and improve quality. Using a photo-analyser, the system could autonomously inspect over 1,000 buns per minute for colour, shape and seed distribution; and continually adjust ovens and process automatically. This resulted in the elimination of thousands of pounds of wasted product per year; shortened speed of production; energy savings and a reduction of manual labour costs.

Another category of intelligent apps are virtual personal advisors or assistants like the oncologist advisor from Memorial Sloan Kettering. This uses IBM Watson on the backend. The premise is that a doctor can’t keep up with all the oncology trial information that is out there. And as a group of doctors get together and discuss new treatments for a cancer patient, they try to pull research from various sources; the advisor is another feed into that committee and they were able to notice unique insights they didn’t know or couldn’t have seen within the doctor’s own human capabilities.

What you have here is a high-value user or use case – the potential death of a patient – and a complex problem with a large amount of data. Unfortunately, much of it is scattered and overwhelming when thinking about the amount of research from oncology trials around the world. And (hopefully) the result is a high-impact treatment plan!

 Collaboration Analytics

Collaboration Analytics is a big piece of the operational initiatives to support next generation workspaces in the digital enterprises. IT Consultancies are writing analytics around Unified Communications that include conferencing, voice and video and messaging.

As part of the next generation of workspaces for tomorrow, it’s important from an operational efficiency perspective to know that the tools deployed are being used – actually, not only that they are used, but by whom, and how often. 1/4 of organisations focus more on the successful implementation of collaboration technology rather than how it’s being adopted by the employee base. How do they know is they’re really getting a return on their investment?

Through collaboration analytics, companies will achieve license optimisation and cost savings and they’ll get a truer sense of adoption patterns and use of collaboration technology in their business. Finally, the value is in the benchmarking. Being able to compare across companies the usage of various collaboration tools will be important. Comparing a 500-person company vs. another similar company will help gauge where that company is among its peers and if more investment is necessary to gain better competitor advantage and innovation. Once we can gain the valuable information of which collaboration tools are being used and where, whether in your traditional meeting rooms at the office or in your rolling office in an autonomous car during the daily commute, organisations will be in a better position to make decisions on investments around these tools to support workspaces for tomorrow.

Coming up in part three I’ll be talking about Digital Technology platforms which are the basic building blocks for digital business, stay tuned!

Footnotes: (2) IDC FutureScape: Worldwide Robotics 2017 Predictions, November 22, 2016 ?(3) IDC FutureScape: Worldwide Healthcare 2017 Predictions, November 15, 2016