Solving the Right Problems with Artificial Intelligence

By Ylan Kazi, Chief Data Officer, Blue Cross Blue Shield of North Dakota

“If I had six hours to chop down a tree, I’d spend the first four hours sharpening the axe.” – Abraham Lincoln

More companies are continuing to embark on their artificial intelligence journeys. Each one is at a different stage in their maturation process. Some are starting from step 1, while others may be midway through. Regardless of where they are at, this journey is fraught with danger and potential missteps. There have been millions in wasted investments, because companies don’t take enough time to think about what problems they want to solve with artificial intelligence.

Starting With “Why”

In my work with larger healthcare companies and consulting with companies outside of healthcare, I’m surprised by how eager these companies are to do artificial intelligence, but that haven’t answered the question of why? There can be the shiny object syndrome, where artificial intelligence is now the next best thing. There is also the danger of thinking that artificial intelligence is the solution to every problem or business challenge that exists in a company (hint: it’s not).

Understanding the why can clarify can start the process of clarifying which business problems to solve. This ensures that you are focused on the problem and then the solution, not a solution in search of a problem. I’ve listed out three why statements below:

Why Statement #1: I want to use artificial intelligence because my competitors are using it.

Why Statement #2: I want to use artificial intelligence to bring a competitive advantage to my business.

Why Statement #3: I want to develop more intelligent products that will increase revenue.

Notice the differences between the statements. The first why statement is vague, and a company embodying this view will get obliterated in the market. The second why statement conveys more specificity but is still vague because it doesn’t answer what the actual competitive advantage could be.

The third why statement includes the right balance of specificity and is simple enough that a large organization could initially get behind it. There is also a clear outcome to knowing that if artificial intelligence is successfully implemented, it should lead to an increase in revenue.

Continuing With “What”

After defining the why, it is important to continue with the what, specifically what are the problems that can be solved using artificial intelligence. Let’s say that you come up with a list of 50 business problems. Out of those 50 problems, there may be only six that would benefit from using artificial intelligence as the solution.

In applying artificial intelligence as a solution to a business problem, there are three key criteria to pay attention to. First, there must be the right systems and processes in place to implement the solution. Your data scientists can make models until they are blue in the face, but if they can’t be implemented, artificial intelligence won’t solve anything. Second, there must be existing high quality data applicable to the business problem, or the ability to acquire this data. Data is the lifeblood of artificial intelligence, and without this, the risk is that the solution will be heavily biased. Third, there needs to be senior executive sponsorship for the business problem that you’re trying to solve with artificial intelligence. This ensures that priorities are aligned and that any roadblocks can be removed along the way.

Finishing With “How”

After knowing why you want to use artificial intelligence and what problems to solve, figuring out how to do it is the last step. If you’ve established the right momentum, the urge could be to immediately hire a data science team and be off to the races. Don’t do this.

If your organization has never implemented artificial intelligence, it is risky to start building a data science team. What I’ve seen happen is data scientists are hired, and then 6-9 months later they end up leaving the team, because they can’t get work done due to a lack of process and subpar data.

In this type of scenario, it would make more sense to hire an outside firm that could take one of the business use cases from idea to implementation. What this does is level up the internal teams and show what it takes to complete an artificial intelligence initiative. During this initiative, gaps will surface, which can present useful learnings. These gaps can include needing to reskill or upskill team members, upgrading the technology stack, and potentially needing to change governance structures.

If your organization is further along in its artificial intelligence journey, it may make sense to create a new data science team or continue to add people to an existing data science team. The key here is being able to show the value that the team is contributing and how adding new team members would increase the value delivered.


By understanding the why, what, and how of your potential artificial intelligence initiatives, you can increase your overall likelihood of success. You may go through this process and realize that you can’t answer the what, which is okay. That may stop you from moving forward and potentially saving millions in misallocated investments that don’t pan out. Consider it a success.

Implementing artificial intelligence in your organization is about solving the right business problems with it. Artificial intelligence is one of many tools that can be used to find value in data, and when applied to the right business problems will get you that 10x or even 100x value for the organization.