Experiments led to increased productivity and reduced costs

In an era where digital transformation and AI integration are converging, the tools we choose can significantly impact our productivity. Powerful tools in the hands of skilled craftspeople can deliver impressive results. A recent project with a longtime client validates this mindset.

A pair of ITX Delivery teams had been working alongside their client counterparts for years; together, they’ve used Miro’s collaboration tool to support their joint activities. When ITX Product Owner Brandon Pierce discovered ways to leverage Miro’s AI Assistant, he started encouraging his teammates to play with it a bit to see if there might be room for it in their shared toolbox.

Initially, the client’s reluctance to bring AI technologies into their product development process constrained the team’s early work. That is, until Brandon’s experiments started to show real promise. Encouragement from these early-stage experiments proved to be the gentle nudge the client needed to explore Miro’s AI Assistance more thoroughly.

Experiments with AI Begin

Beginning in early 2024, our client began to evaluate the new Miro AI Assistant tool; their objective was to determine its potential benefits and return on investment. The AI Assistant was relatively new to the market, but it built off the Miro mission to enhance team collaboration and productivity. Known for its deep integration with Miro’s collaboration suite, this AI tool promised to bring significant advancements to the team’s sprint-based workflows.

As the client began to explore the Miro tool, they quickly realized that their earlier resistance to AI tools generally limited their progress. So the client called on a group of ITX Miro power users to lead the experiment’s next phase.

Taking a Methodical Approach to Exploring AI

Our early-stage observations led us to believe that Miro’s AI Assistant would deliver significant – and immediate – business value especially with respect to streamlining a variety of sprint-related tasks. Now we had to prove it; that is, we had to clearly demonstrate the tool’s ability to deliver the quality we expected, generate efficiencies, and eliminate unnecessary costs.

ITX power users hypothesized that the Miro AI Assistant could help the team prepare sprint-oriented activities like planning, dependency management, standups, and overall readiness. They were also hopeful – though less confident – that the Assistant could:

• Automate routine tasks and provide suggestions in real time, allowing subject-matter experts to focus their energies on higher-level tasks;
• Expand on team member ideas, drilling into insights in ways that enriched team discussions and planning sessions
• Facilitate the writing of user stories and acceptance criteria; and
• Improve the accuracy of these and other outputs through the use of the Assistant’s capabilities around bias detection and language translation – which was particularly important for our team as it allowed native Spanish speakers to express their ideas in their own language and then translated them to ensure clarity and understanding.

Our objective was clear: rigorously test the Miro Assistant to determine whether it was capable of providing meaningful business value to our client.

Context and Challenges

Before diving headlong into the experiment, we took a close look at the entire project – not only the opportunities that AI and machine learning provide, but also the hurdles we needed to overcome to realize AI’s well-advertised potential. As we performed our Discovery tasks, we learned the following:

• Security Concerns. The client’s initial resistance to integrating AI tools into their processes was due to security concerns about the risk of exposing company details and proprietary information being exposed.
• Limited Experience with AI. Our client had limited experience with AI tools, so part of the ITX mission was to use our experience to help them navigate uncharted territory.
• New Market Offering. At the time, the Miro AI Assistant was a fresh entrant in the market, which added an element of uncertainty regarding its efficacy and integration.
• Resource Concerns. Understandably, our client shared concerns about integrating a new technology into their product development operations and building proficiency with it.
• Strategic Importance. Integration of new tools and skill development are significant investments. ITX’s challenge was not only to help our client appreciate the short-term dividends, but also to understand the broader, long-term strategic implications.

Our Approach

After years together as technology partners, we’ve embraced our client’s general operating philosophy, their tolerance for risk, and their day-to-day ways of working. We developed a structured, methodical approach to reflect their preferred style.

Understanding Capabilities and Limitations. Trust is hard-earned over time, but easily lost and difficult to regain. That’s why we began our experiment by engaging in information-sharing sessions with Miro to fully grasp the AI Assistant’s capabilities and limitations. We shared this foundational knowledge with our client to, which not only enhanced our transparency, but was crucial in conceptualizing how Miro Assistant could be applied to their operations.

Experimentation and Testing. Next, we conducted a series of experiments on the new AI features. We created a shared dashboard to track our progress and facilitate communication among team members running separate tests. At the conclusion of each week’s test, the project team – including representation from the client, from Miro (who helped facilitate the project for our client), and ITX – gathered to discuss their findings and introduced new opportunities to improve our current activities. These discussions also benefited Miro’s product teams, who applied our suggestions to improve AI Assistant for the long term. This collaborative approach ensured that our learnings and anecdotal insights were consolidated in one best place, able to be analyzed by all members of the team.

Applying Learnings to Team Operations. Across the team, we could feel the momentum building. Armed with new data-driven insights, we confidently applied the AI Assistant directly to team planning sessions. We observed the emergence of new process efficiencies; we noted increases in the quality of our outputs; and we were able to identify and analyze unexpected results. This phase was critical in translating experimental data into practical, actionable improvements.

Results, Impact, and Next Steps

The outcomes of our experiment were illuminating. Here’s a breakdown of what we learned:

• Strategic Value. Our most significant finding was the strategic value of investing in Miro Assistant. Because our experiments with AI took place in a controlled environment, client teams were able to move at their own pace and developed a deeper understanding of its potential and limitations. ITX’s methodical approach was instrumental in building the team’s confidence and competence with the Assistant.
• Innovative Use Cases. We identified several innovative applications that went beyond our initial expectations. For instance, the AI’s ability to expand on our ideas and uncover biases provided valuable insights that enriched team discussions and planning sessions.
• Enhanced Quality. The accuracy of the team’s outputs improved notably. Super cool features, like language translation and bias detection, helped us produce higher-quality work by simplifying communication and improving visual aids. This, in turn, facilitated clearer and more effective planning and execution.
• Efficiency Gains. As we thought, the Miro Assistant proved to be a powerful tool in streamlining our sprint tasks. With our dashboard, we tracked significant time savings in processes such as story writing, review, and sprint planning. The AI’s ability to automate routine tasks and provide real-time suggestions allowed teams to focus more on strategic decision-making.

Recommendations and Future Direction

Based on our findings, ITX presented several recommendations for future iterations of Miro Assistant, its implementation, and beyond:

• Refinement of Features. Continue to perform Discovery activities; gather feedback from users to drive the refinement of the tool’s capabilities. This best practice will not only enhance the team’s experience with the AI Assistant, but it will also present benefits in areas like customization and integration with other tools and technologies.
• Training and Support. Provide regular training and support so that team members can streamline the adoption of new AI tools to maximize their business value. Remember, powerful tools in the hands of the unskilled can yield undesirable, even disastrous results.
• Scalability. Plan future integrations of the Miro Assistant (as well as other AI tools) to provide the best chance of scales effectively with different team sizes and workflows will be crucial as more organizations consider integrating AI into their processes.

Conclusion

The primary goal of experimentation is to learn. It’s an added bonus when results align with your hypothesis. And that’s what happened here, an experience made even better because ITXers collaborated so closely with our client counterparts.

By carefully testing the AI Assistant and exploring opportunities beyond our initial experiments, we demonstrated that Miro’s AI Assistant is not just a powerful tool, but also a strategic asset. The business value we were able to show – in terms of improvements in quality and productivity – underscores the importance of thoughtful experimentation that embraces innovation.

---