Are you responsible for kicking-off AI projects in your organization? And do you need to explain your AI use case to management?
In a previous post, I have written about how I spoon-feed management step by step. Once they understand the gist of AI, and I have an AI use case to pursue internally. It is time for setting up that first meeting to introduce my use case. In this meeting, I keep it simple, I have four key messages I use to explain the use case.
Message 1: The solution in one sentence
I write down the solution in one sentence. What is it exactly that my solution will do? Going through the process of writing it down one sentence helps me to find the right explanation and make it easier to get them on board. And I make sure to focus on the business side of things.
A good example: “Automatic acceptance/rejection of credit applications”
A bad example: “A logistic regression for prediction of paying back loans”
Just think of what will happen in the mind of your manager. In the first example (s)he will think “Great. I can save costs”. In the second example (s)he will think “Mm, interesting”.
Message 2: The benefit of this AI project
In the second step, I speak about the benefits of this project. Benefits can be hard factors such as time saved, an increase in ROI, cost savings, or a reduction of a certain risk. Sometimes the benefits are soft, and that is OK too. To make my management really interested in a use case, I make sure to mention the KPIs they have linked to their bonus.
Message 3: The data this project needs
AI projects need data, but my management may not fully understand this or have not thought that far ahead. In this step, I tell them what data I will need, where it comes from, and how easy or difficult it will be to get this. This really brings my case to life. Once they understand what will go into the model it starts to become tangible for the management as well. I find this approach sparks concrete discussions and helps them to come up with other ideas I may be able to include.
Message 4: The key success factors
We have all been there, too high expectations from management. In most AI projects, there will be roadblocks down the line. I think about these roadblocks ahead and turn them around into key success factors. This first phase of a project is my time to set the right expectations for the management. It will help me tremendously down the line when problems arise that I have anticipated upfront.
These are the four key messages I bring across when first initiating an AI project with management. Most of the time one meeting is not enough and I go through a cycle of a few interactions until the message has landed. When the message has landed, then it’s time for the next step.
Be patient, it may take longer than you anticipated.
About me: I am an Analytics Consultant and Director of Studies for “AI Management” at a local business school. I am on a mission to help organizations generating business value with AI and creating an environment in which Data Scientists can thrive.