The world of Artificial Intelligence has boomed in the last two years. Companies throughout the world have looked for ways to integrate AI into their products in an effort to improve the workflows of their users.
Our goal at ServiceMax was to design a tool that helps personas in the field service world get their work done faster. The challenge they faced was that legacy systems and hard to obtain information often hindered their ability to move forward on the actual nuts and bolts of their jobs. Wouldn’t it be better if they had something that was available to them, at the ready, in a familiar location that could help them cut through the fluff?
As the lead designer, it was my job to work closely with the Product Manager and Engineering team to deliver an AI assistant that would help keep the world running.
Challenges we faced
Implementing a new technology such as AI is a road that is filled with lots of lessons in product development. Here’s a few of the challenges we faced and what we learned from them.
Determining the unknown
Since AI technologies were so new and changing on a day to day basis, it was difficult for our team to know what we could capably do. Early impressions was that AIs could do anything as long as you enabled them correctly but what did that mean for our current technologies and how could that incorporate with Salesforce, the platform our product is built on?
While the product team brainstormed potential use cases that would help our customers, we also had to come up with designs and check their feasibility against the AI tools and the current tools we had in place.
Making decisions on what would get built
We ultimately had to make decisions on what use cases we wanted to design and develop. Our usual process is to leverage what we know about customer needs and delve into their pain points. While this was similar to how we determined what we would focus on, there were also other business goals to consider.
Leveraging current technologies
Our product is made on Salesforce. Thus, our customer’s data lives in Salesforce. This meant that the tool we used would have to be able to consume data that was in Salesforce and often formatted is various different ways by our customers.
Design iterations and working closely with engineers
The nature of the nebulous situation meant that there would be many design iterations. While designing, I would work closely with the Product Managers to determine whether we were meeting our requirements. After we felt our designs were in a good spot, we’d share them with engineers (who were researching these AI technologies). Often times, our requirements had to be adjusted given the new learnings the team had uncovered. This meant that designs had to be changed on the fly and we were working in a quick iterative process.
Beta Program
This feature is currently in a beta program with select customers providing feedback on their experience. From this program, we look to improve the features the Copilot provides.
