Designing for Trustworthy Automation
Lessons from creating the UX of a virtual assistant. Employers wear many hats, and it can be difficult to fit hiring into their schedule. Besides, finding new employees and understanding job market trends isn’t necessarily their area of expertise. Part of Indeed’s work is to make the hiring process more efficient for employers. My team supports these improvements by developing automation tools, and recently we designed a virtual recruiting assistant to automate part of the hiring process. Our virtual recruiting assistant included three key features: Data-informed improvements to the job posts to attract the job seekers with the right qualifications. A daily report of promising candidates for employers to review. Interview auto-scheduler based on employers’ availability. While we knew that our tool could help employers make effective hires, we ran into many challenges in getting employers to use it. After 12 months, we ended up retiring the virtual recruiting assistant, but that doesn’t mean it failed. There’s so much that we learned during that year. Today, that knowledge continues to inform new automation projects. Your team might be able to learn from our challenges, the adjustments we made to overcome them, and the wins we had. Come follow along. Help employers understand automation Automation is a relatively new concept for employers, and many of them are skeptical of its advantages. They simply aren’t always convinced they need automated tools. When employers work with a human recruiter, they can meet to ensure the recruiter understands their needs and ask for specific recommendations. With an automated tool, there are no humans to talk to. In our user research studies, employers mentioned that they prefer a personal touch. A quote from our research findings says: “…employers have inherently different expectations from a human recruiter versus AI, and are less forgiving of AI’s mistakes or annoyances.” Helping employers understand what automation does and what outcomes they can expect can help your tool gain momentum in the face of these challenges. Here’s what I recommend. Show, don’t tell Research showed our team that even when employers understand automation, they’re reluctant to use it because it feels impersonal. We also learned that employers may feel automation puts key decisions out of their control, which highlighted a trust issue. So our first challenges were to earn that trust, help employers understand how the automation tool worked, and effectively explain how it could benefit them. We knew from our data that automation helps employers make hires, but automation is like going to the dentist—people understand that it’s useful, but they don’t necessarily enjoy it. Employers wanted to know how automation would change the hiring experience, but automation is a complex process that requires lots of explanation. Adding new features increased the amount of explaining we had to do, and employers complained of information overload, misunderstanding, and fatigue. They gave up on the automation because too much information made it harder to understand. We tried several ways to tell employers how our automation tool worked. But these solutions weren’t the most effective approach. When we started showing how the tool worked with illustrations—rather than explaining it with text—we increased the adoption of the paid features and helped employers get closer to making a hire. We also saw a significant increase in jobs with interviews confirmed by job seekers. Automation is like going to the dentist. » Read More
Like to keep reading?
This article first appeared on indeed.design. If you'd like to continue this story, follow the white rabbit.