Every great idea starts with a question.
What if we were to…?
How can we…?
Why don’t we…?
Sometimes, the problem (especially if you’re not a formally trained data scientist) is knowing what question to ask. And more importantly how to ask it. That’s where design thinking comes in. A good question usually cascades into more questions that result in a better answer and a better outcome. But how does and how should our thinking change as automation and artificial intelligence become more prevalent and machines start to make more decisions? Let’s start from square one…
So, what is design thinking again?
All great innovators in literature, art, music, science, engineering, and business have practiced design thinking. Chances are, you’ve used it at some stage in your career without even knowing it. It’s a method in ideation and development that describes a human-centered, iterative design process consisting of five steps: empathize, define, ideate, prototype, and test. It’s often compared to, and aligned with, agile development practices. Design thinking has been used for years to tackle problems that are ill-defined or unknown, which makes it the perfect methodology for data scientists and business analysts tackling big business problems with little direction and an abundance of data.
Design thinking at its core has remained the same in principal and objective, but it has been forced to evolve over the years as new developments in business strategy and technology have emerged to influence our thinking and decisions. For those of us working with data, the continued advancements of machine learning (ML) and artificial intelligence (AI) technology will undoubtedly impact how we uncover insights for our organizations, but one thing is for certain: Humans will always be a lynchpin in design thinking, and design thinking will always be a lynchpin in finding the best answers.
Getting to (Wh)Y: Design thinking to uncover what machines cannot
If you read the headlines today, everyone is afraid that machines are going to take their jobs. It’s true. Automation and technology are changing the modern-day workforce, but humans are still—and will continue to be—a critical part of all technology and automation. Technology and automation are simply enhancing human’s ability to make a decision and move faster, but they’ll never replace the need for human intuition and experience – two critical facets of design thinking.
Let’s say you were part of a data science team, and you’re asked to solve a key scheduling problem for a hospital. The problem seems straightforward: Predict how many nurses would be needed each hour, by day of the week, so that an optimal schedule can be produced.
After asking a few questions about key constraints, the data science team is thrilled to engage in this endeavor, as it sounds like the perfect use of data science. Forecasting, after all, is a key tool in the data science arsenal. The executives kicking the project off share that a shift can be eight hours or 12 hours, and that each full-time nurse should be scheduled at 40 hours per week, and each part-time nurse scheduled between four and 20 hours per week.
So what would your team do? Brainstorm what data might be useful. Look for data that is available. Could the weather be a contributor to the load? Some think a full moon might be a driver. Holidays? And do we have the historical loading data to build our model with?
But here’s what happens frequently: The team gets the data, builds the models and goes to implement it—and then finds out that it just doesn’t work.
How could this happen?
Because all the answers that the team needed to be successful were not just in the data. Well, in this case, if you visit the hospital and watch what happens when the loading spikes, you quickly see that there are some nurses that are on-call. They come in just for the spikes and really change the way the loads are handled. And this wasn’t shared with you in the communication of the constraints—but it is completely obvious when you live the experience with the team.
You also learn about what happens when nurses get sick or are unable to come in for their shift. Are we predicting these shortages in our approach?
When do the nurses need their schedule? How far in advance do we need to produce the forecast? These are all items that are very evident when we “walk the process” and meet the participants, and ultimately, understand the “why.”
And after living the experience, we may find that the larger issue is to predict when a spike is one hour away, as this is the lead-time the on-call nurses need to smooth the load. Once this issue is solved, scheduling can be handled very differently than without this taken care of.
This example is the heart of the design thinking methodology and showcases exactly why humans need to be involved at almost every step of problem-solving, from ideation to testing. Machines cannot empathize (at least not yet…), and often times it’s necessary to physically experience the situation—use your five senses and really understanding the problem firsthand. With this added insight, the lens to view each problem is sharpened, and the ability to implement transformational results is increased.
With great problem solving comes great thinking and while machines are showing that they are able to think faster, it does not mean that any of that thought is critical or structurally sound. Human intervention is still a necessity in any problem-solving scenario. The existence and rapid growth of AI and ML technology cannot, and should not, threaten a human’s ability to tackle the world’s greatest challenges and answer the most complex questions using design thinking. It can only enhance it.