Tijl van der Giesen 621445
Ruler is a tool that helps organisations keep track of regulatory developments. After a year of working with Ruler, I knew the platform well, but this project showed me how much of its editorial work I had never explored. I suggested it for our AI Strategy Lab project because I found the proposed workflow innovative and feasible, and it made me personally invested in the outcome. I expected that AI could make the preparation of regulatory updates considerably easier.
My role at Ruler had been similar to that of a junior project manager. I created client accounts, configured profiles and gave demonstrations. Because I had no legal background, I had not worked directly on the publications myself. This project helped me understand how the information provided to clients was prepared. I enjoyed learning about this, and it made me appreciate the tool even more.
My main contributions were introducing the workflow idea and developing the problem and opportunity, business model and future-market test. I also contributed to the evaluation, spoke with consultants about preparation time and hourly rates, and selected articles from Ruler for testing. In the final days, I made my slides and carried out final checks. I also reorganised the deck to explain the current workflow on a separate slide before introducing our proposal. This improved the storyline and helped us show why the change mattered. The whole idea seemed pretty straightforward and easy to execute, but I underestimated how much time a complete consultation required. Even with AI, reviewing and correcting the output required more effort than I had expected.
Initially, I wanted to explore an agentic AI system that could handle almost the entire process independently. Conversations with consultants and the lectures on combining expertise with AI helped me to develop a more realistic workflow that could actually be feasible. Someone still needed to judge whether an update was relevant, accurate and suitable for clients. We therefore proposed AI preparation followed by consultant review, correction and approval. We defined success as reducing the total time and effort needed to produce an approved package while maintaining quality.
My teammates mainly built the prototype, which came along pretty quickly and worked surprisingly well in our test. I contributed ideas about displaying alerts so consultants could clearly see when something still required attention. Our tool was about the cooperation between consultants and AI, and this made it easier for the reviewers to see which draft needed more attention and work.
Our five exploratory cases suggested potential time savings, even though every draft needed editing. Finding a realistic comparison was difficult because the consultants I spoke with had limited time to help. We compared their estimates of manual effort with AI-assisted times recorded by a team member. The results were encouraging and suggested that AI could save time. A pilot with professional consultants would help us confirm whether these savings hold in practice, measuring both workflows in the same way and including the time needed for review and correction.
Collaboration went smoothly. Because my teammates were not familiar with Ruler, I had to explain certain things which forced me to think through the process more carefully. I learned more about the company and business model because of the conversations and discussion we had. My main discussion was about the business model, concerning Route A, fewer billed consultant hours, and Route B, using released capacity to expand coverage. From my experience, Ruler already covered almost every relevant publication, so I questioned whether doing more would generate value and if there was enough demand. We chose Route A as a clearer financial case from Ruler’s perspective. This helped me understand value creation and value capture in practice. Saving time could make the workflow more efficient, but Ruler would only benefit financially, under Route A, if those savings translated into fewer billed consultant hours. The consultants would still play an essential role in ensuring the quality of the final output.
The future-market test also helped me better understand a key message from the lectures. Other organisations could use similar AI, so the technology alone would offer limited differentiation. Ruler’s existing platform, infrastructure and consultants could help turn it into a useful service. This helped me understand how complementary assets, such as existing systems and expertise, can make AI more valuable. It also made me more interested in how companies can use AI to build on their existing strengths.
At Demo Day, other students described our idea as innovative and feasible. I was also asked whether we had considered the consultants’ perspective. I explained that our financial analysis focused on Ruler, and I still think that was a reasonable scope for this assignment. Making the idea work in practice would also depend on the consultants. Fewer billed hours could mean less income for them, which might affect their willingness to use the new workflow. Their support would be important for achieving the expected savings, so Ruler would need to discuss these concerns with them before implementation.
If I could restart the project, I would really think about the comparison before developing the prototype. First, I would clearly map all the tasks of the consultant: reading, assessing relevance, finding links and drafting. I would then try to arrange a small test that fitted the consultants’ schedules, comparing both workflows using the same publications and quality requirements. This would include the time needed to review and correct the output. Planning the test earlier would help us identify where AI saves time and where it adds work. Our first test was a useful starting point, and this approach would give us more confidence in the results.
I am satisfied with the final product and believe a similar workflow could be implemented at Ruler. I will try to pay more attention to how AI fits into the work people actually do, whether it saves time and who benefits from it. Once more, I really enjoyed working on Ruler because it gave me a chance to explore an improvement for a platform I already knew from experience. I still keep in touch with people at Ruler, and I would like to discuss this project with them and hear their thoughts on the proposal.
