The study resulted in improved data ingestion processes, resulting in a 20% increase in customer satisfaction, giving our company a competitive edge in the market. This study also achieved a 50% reduction in time to market by orchestrating sense-making processes, identifying critical Jobs To Be Done, mitigating risk, and swiftly testing cost-effective artifacts.
How might we create an experience that speeds up the amount of time an agent needs to start an estimate for an assignment and move on to the next assignment.
Based on previous research, and while leveraging our in-house subject matter experts, we hypothesized that field appraisers will follow a flow around the vehicle, similar to a clock face, stopping every few steps to take a photo. We also hypothesized that they would take additional photos and want to take notes when they come across unrelated prior damage or the damage itself. Two design colleagues and I began ideating with artifacts and assembled a prototype that we tested with customer end users.
I assembled a remote moderated test plan, assembled a team of 3 (I moderated each session while a colleague observed.) and scheduled evaluative research sessions with 5 carrier field staff from two partnering customers. All sessions were remote as covid levels were unsafe for in person study sessions at the time.
UserTesting is the tool of choice for all qualitative generative research, moderated evaluative research, and unmoderated usefulness testing.
Our evaluative study helped us validate design decisions, and identify constraints that resulted in more ideation. In parallel, there was an evolving technological advancement from our data science foundry and our AI model. This produced the addition of AI predictions though machine learning. Photos submitted through the AI model would be scanned and the output would be a predicted list of parts and labor hours. For instance, if you uploaded a photo of a damaged vehicle door from in the field, the AI model would predict if the door needs repair, replacement, or needs removal and inspection. In addition, it would predict how many labor hours would be expected to complete the repair.
We held a 3 day design sprint and gathered "How Might We" feedback based on our SME interviews, and spent another two days designing artifacts to add to our prototype.
Day 1: Subject Matter Expert Interviews
To prep for the sprint, we hypothesized and were able to narrow our curiosities down to four main topics. We then brought in our internal subject matter experts and spent 4 hours deep diving into our hypothesis and gaining a perspective into what would be successful outcomes.
Day 2: Prototype
Our second day was dedicated to ideation. We took the data we learned from day one and applied it visually.
Day 3: Internal Evaluative Testing