Context

• Drones, especially industrial drones, generate terabytes of data and will be a major wave in any industry where humans have to look at physical infrastructure and make decisions on it.
• Intel wants to own those data pipelines.
• They partnered with Delair, a French aviation startup, to rebuild from the ground up a platform for analyzing data gathered from industrial drones, branded as the Intel Insight platform.
• Intel hired a small UX agency--on which I was the project lead--initially just to design what the Intel CEO would present at a drone industry conference, but because of the success of that demo, retained us to simplify the product and make it usable.

Challenges

• Complex team structure. Intel had one set of priorities; the French startup had another set of priorities. Either group could stall project progress and neither had experience building software in a lean or customer-centric way. Both corporate cultures tended towards top-down, waterfall development. The teams were also separated by language, physical distance, corporate culture, and time zone.
• Technical limitations. Data sets from drone flights can be hundreds of gigabytes or more. This software performed analytics on these massive data sets in the cloud. In addition, the algorithms needed more training data, so we didn't have a clear sense of the ways in which they could fail and, consequently, how to make systems degrade gracefully for the user.
• Time to a decision. The team was fortunate to have line-of-sight to the C-Suite at Intel, but that meant that the team sometimes felt disempowered to take risks without first checking with layers of management.

Approach

• ‍Ask why, a lot. The French engineers had more devotion to the problem than anyone else. Which meant they also had strong opinions that weren't always customer backed. Asking them what they liked about their proposed solution unpacked their motivations and helped me frame solutions in the language they wanted to hear.‍
• Meet complexity with structure. Daily standups, frequent updates to the sitemap, solicit feedback early from key stakeholders.‍
• State assumptions explicitly. Echoing what I thought I heard from key stakeholders helped penetrate cultural and language barriers.
• ‍Focus on differentiation. Other drone analytics software exists, but as we reviewed it, the software was often bloated, hard to learn, and work products were difficult to share with other collagues. We knew that wouldn't work with target users' workflow. So we focused on simplicity and enabling communication inside the platform.‍
• Generate many solutions. Drone analytics is a very new spaceIn a space as new as drone analytics, the competitive space is new and rapidly evolving. The technical constraints were also not obvious in advance. So, to drive the right conversations and find unexpected solutions, I tried to bring 3+ solutions to a problem to the table at every juncture, and often many more.

‍

Outcomes

• I left the team with a set of completed designs for the beta launch and wireframes for the general market release. 
• In addition to redesigning the UI in its entirety, I advocated for and got approval for several major shifts in the product: focusing on communication features and consolidating information from multiple flights into one project.

‍

Takeaways

• Conduct frequent retrospectives. We didn’t reflect on what we were doing well and what we could be doing better often enough.
• Let people sit with designs. To empower better critiques, I needed to give people a gestation period to sit with a design and consider it from multiple angles. This was especially true of technical feasibility conversations.
• Protect space for open-ended conversations. I spoke to the dev team in France every day. Protecting that space ensured that we didn’t let misunderstandings or questions fester.