As with any new transformational technology, business leaders often rush toward whatever new "shiny object" promises to streamline their business. For artificial intelligence (A.I.), this was especially true during 2020, as a recent survey found 43 percent of businesses around the world were accelerating their A.I. initiatives in response to the pandemic.

Unfortunately, many of these businesses rushed to integrate A.I. into their workings without stopping to ask who, how, and why. As companies look to take advantage of the business insights and other benefits A.I. can provide, it's important they don't attempt to put square pegs in round holes.

A.I. can appear magical, but it isn't magic. Bad algorithms yield bad results. While investment and experimentation are extremely important, the biggest and most common strategic mistake companies make when exploring A.I. is failing to define a clear use case and desired outcomes with a clear, quantifiable metric for the technology in the first place.

To solve this problem at my workplace, we decided to turn to the principles of design thinking. A human-centered approach to A.I. starts with who will be consuming the A.I., how they will be consuming it, and why the A.I. is even needed. This starts with thinking critically about the problems your business is facing, framing those challenges in ways that are potentially solvable by A.I., and then identifying and refining use cases that are critical to your business goals.

With a data-driven and human-centric approach, we as business leaders can design A.I. that successfully connects every strategic data and A.I. initiative to the defined business objectives of a company. If you are interested in investigating how A.I. might be helpful to your own organization, I'd encourage you to follow a similar approach.

1. Set intent

Many companies don't really have a clear idea of what they hope to get out of A.I. beyond some vague notion of efficiency. That's why it's important to refine your intents by spending some time uncovering the targeted A.I. business opportunities that exist within your current business strategy. Are you trying to keep workers safe? Keep customers happy? Begin with a clear intent that is grounded in your core business objectives.

2. Identify

Once you have determined your overall objective for implementing A.I., you can then define the use cases and the types of A.I. solutions needed by the users and that will eventually be integrated into your infrastructure. A.I. is rapidly advancing in numerous fields, from computer vision that determines what's in an image to the natural language processing A.I. that you find in chatbots and virtual assistants. What are the ways these applications can advance the intentions you outlined?

3. Evaluate

The evaluation stage involves figuring out what data you need to make the use cases you've identified effective. Different kinds of teams focus on different priorities and different sets of numbers, meaning that most industry data is siloed to some degree. To implement successful use cases through A.I., you need to ensure your A.I. is being fed accurate, clean data that draws from your entire organization.

4. Plan

The last step of the design thinking approach focuses on setting concrete actions by using statements of intent as a guide for the technical implementation. The goal is to help customers operationalize A.I. through the business by connecting every solution to the defined A.I. strategy.

Critically, an implementation strategy must account for user trust: How will your customers or clients react to your organization using data in this way? How can consumers and the public know that your implementation of A.I. is explainable and trustworthy?

Designing a successful A.I. strategy is also about who has a seat at the table. It's important that businesses include diverse voices and the right stakeholders at each stage of the process.

In my workplace's approach, the strategy-setting sessions are attended by the senior business executives who set the intent, define information types, craft business hypotheses, identify use cases, and infuse company ethics into the strategy. The technical sessions invite data scientists, designers, and developers to come together to transfer the intents set in the strategy session into a detailed strategy, defining the use cases, evaluating the data, and planning the execution. Throughout each exercise, visual storytelling, images, and graphics are used to help ensure that, though they come from different fields, everyone involved gets a chance to speak the same language.

The most common takeaways? Frequently when I work with clients, their aha! moment comes during the "evaluate" phase. All too often businesses believe they already have all the data they need to run whatever A.I. models they want. This is rarely, if ever, the case.

For example, one client in the financial services industry wanted to develop an A.I. solution that would help quicken the economic recovery of small businesses impacted by the pandemic. But, when assessing the data needed to create value for the selected users, the team realized for the very first time that their data was disorganized, siloed, or not usable. Before starting to implement a reliable model, you need to fix data collection, infrastructure, and platform issues that hinder the development of trustworthy A.I.

There is little question A.I. is already transforming business today. From health care organizations using natural language processing to help process Covid-19 related queries to financial services companies using A.I. to parse tedious compliance documents, A.I. early adopters are continuing to develop new use cases by the dozens. But what these successful implementations all have in common is a clear intent and plans that connect the advantages of A.I. with a business's main priorities.