The AI Strategy Lab was an interesting assignment in which I worked with three other students on an AI governance assistant (Slipstream). Slipstream works as a silent observer that catches missed commitments before they turn into project delays. Slipstream gathers communication from different channels like Outlook, Teams, and Slack and provides the project manager with daily updates and a dashboard.
This idea came from a brainstorm, where we all came up with 1 or 2 proposals that fitted the criteria of the assignment, and voted which suited best. I came up with an idea on using genAI to make the process of graphic design more efficient and cost effective, without losing human creativity. After the voting round, we chose to continue with the business idea that resulted in Slipstream. After submitting our proposal, we received some feedback, which we applied before moving on with the project. The feedback we received was really helpful, as it guided us towards better aligning our project idea with the assignment requirements.
My contribution and role during the project can be split into four main parts. Firstly, I rewrote our proposal based on the feedback and added additional sources, which ensured that we all had the same understanding of the project. Secondly, I have made a proposal for the structure of the final slide deck. After some quick brainstorming and small adjustments, we used this structure for the final slide deck. This has helped our team, as it has enabled everyone to get started on organising their topics. After the deck structure was done, I focussed on the business model, the economics and pricing, and the ecosystem analysis of Slipstream. Finally, on the day of the deadline, we as a team wrapped everything up and created the slide deck on Canva. Canva was a platform I had no experience with, but luckily it was easy to use, and I am happy with the design of the final deck of slides.
The hardest decision during our project was about how to shape our dashboard. Although I not really contributed to the dashboard, I was involved in deciding on the desired outcome. As no one in our team had experience with creating dashboards (even with AI), we had to find a way to make sure our dashboard looked nice while also functioning. Luckily, we discovered that Claude was able to meet both of these requirements. One hard decision in the part for which I was responsible involved the business plan and the economics behind it. I found it challenging to develop the business model and the economics behind it, without having concrete data to support my decisions. The structure of the assignment was also challenging, as I was not used to working toward a slide deck as the end goal. During my bachelor’s degree, I was used to writing mainly reports or essays. The limited space in a slide deck made it tricky to clearly express everything I wanted to say.
What surprised me during the project was how Slipstream performed in comparison to our baseline, which were two practicing project managers. Besides the large difference in review time between Slipstream and the project managers, Slipstream was also able to find significantly more signals. A thing that not particularly failed but didn’t meet my expectation was that Slipstream was not capable of outperforming project managers in identifying false alarms. An important note to all of this, is of course that Slipstream is only review on small, controlled set of communication threads, and that the real world is much more complicated.
The collaboration within our team was a little challenging at the start because we did not know each other very well. We all came from different backgrounds and had different preferences about ways of working and communicating. However, these small problems only occurred at the start of the project. As the project progressed, we got to know each other better and this helped us to achieve the final result. When evaluating the collaboration, I also discovered some aspects from which I have learnt. For example, I don’t usually prefer meetings if I don’t believe it is necessary. But I discovered that meetings, or meeting to work together on the project could also improve the efficiency as it ensures you’re on the same page more effectively than if you were to send a text message about your findings.
The project, and the whole course, made me realize that AI is capable and useful in much more things than I imagined. What really surprised me was that it was pretty easy to create a dashboard without having that much of a technical background when one of my teammates showed the process of building the dashboard. The demo day also gave really good insight into the capabilities of AI in business strategy, even though the lecture hall was not really suitable for the format of the demo day.
