AI is going to revolutionize most everything. AI will give you a much more efficient and reliable workforce. AI will save your time. And overall, AI will allow people to do their jobs better and produce more with less effort. I felt this before we started the project, and I feel it now. The difference the project made for me was to provide clarity around where this is true. It is true if there is a well-defined “job” for the AI, if the AI fits into a viable business model, and if the AI’s performance can be measured relative to current methods. Below is a reflection on the members of the team and the choices we made. Below are the challenges we faced. And finally, below is a discussion about how my views have evolved regarding both AI and strategy.
I was responsible for being the critical voice within the group; keeping track of the narrative throughout the process; questioning the components of the project including the user of the product, the problems we were solving for the users, and whether all aspects of the argument fit together; maintaining consistency in telling the story from problem through economics and evaluation. In addition, I created the dashboard for the prototype which was then presented by me; assisted the team in creating a Canva presentation to support our arguments; developed a name and logo for Slipstream which would represent the product. I utilized an AI Assistant (Claude) to create some of the dashboard elements and assist in developing various logo options. I determined what had to be shown on the dashboard; reviewed and iterated on Claude’s input; ultimately chose which option to pursue.
One of the toughest decisions was how to charge for the product. As a team, we came to an agreement on the pricing model. Our discussion of the pricing model linked the value provided to the customer with the cost to provide that value. Thus, it made the economic feasibility clearer than before. I was impressed at how difficult it is to develop a solution (especially one based on AI) that is both reliable and easy-to-use versus one that simply demonstrates functionality. A tool that provides risk alerts to project managers is of little help if the project manager cannot rely upon the tool’s assessments and ignores them due to poor performance.
While the theoretical aspects discussed in class exceeded my expectations, the create, deliver and capture framework proved very practical. It allowed the team to organize the development of the solution and make it more cohesive and credible. Without it, I believe we would have developed a series of features rather than a complete business model. It provided a common framework for addressing questions related to how value was added; how that value reached customers; and how we intended to profit from the delivery of that value.
The feedback we received was a turning-point. The feedback indicated that while our primary testing effort focused on whether stakeholders would adopt an AI system that automatically read emails, Teams, Slack and other project-related communication; privacy and trust were significant issues but did not demonstrate whether the system was effective at identifying project risks. The feedback further recommended developing a controlled environment using a sample of project-related email communications, meeting minutes and task lists and then compare the output of the AI against the existing project management processes regarding missed commitments identified; false alarms generated; time saved; and usability.
We transitioned from “Will people adopt this?” to “Does it really work compared to what they currently do?” While we continued to address privacy and trust as part of the overall governance consideration for the project, these issues were no longer the sole measures of success. I learned that evidence must evaluate your claims and that a topic can appear important without being the correct issue to measure.
Due to differing schedules, lectures and work competed for everyone’s attention. However, we were able to schedule regular meetings and keep them brief and productive. Each meeting allowed us to discuss what we were working on, why, and how, allowing us to work independently prior to our next scheduled meeting yet maintain an integrated outcome.
This project demonstrated that even though AI technology continues to evolve, ultimately it relies on human judgment and decision-making. For example, in our application, the project manager remains in control, while the AI functions as a supporting agent. With this role, there exists the potential for AI to significantly augment a project manager’s workflow. Specifically, it will save time and resources previously spent performing redundant tasks, enhance visibility into project status and identify potential risks, misunderstandings and missed elements thereby allowing the project manager to focus on key items and deliver higher quality projects.
Additionally, it has influenced my way of thinking about strategic planning. Asking “Can AI perform this function?” is only the first step. Ultimately, the more difficult questions include; Who gets credit for providing value? How does one defend their solution over competitors? What happens when both customers and competitors also have competent AI systems? Creating a worthwhile use of AI requires more than just demonstrating capabilities. It requires a viable business model, empirical evidence and clearly defined roles for humans. When I began the course I thought primarily about what AI could accomplish. I now conclude that I should think more about what makes an AI-based idea worthy of development.
