Fountain is an ATS (Applicant Tracking System) recruiting platform that helps companies hire their workers for high volume hiring. Over the last few years, we’ve expanded our customer base and product offerings to help more enterprise customers in various industries find their frontline workers.
In 2021, we launched an AI chatbot feature called Fountain AI with a goal to answer applicant questions the moment they arise from the customer’s career site and jobs directory page to help customers convert interested candidates into great hires.
Increase Hiring Speed
Personalize the applicant experience
Be efficient
Challenge
The biggest challenge was that this engineering-driven initiative had minimal design input at the time of the launch, resulting in poor UX considerations. After launch, the chatbot feature had low adoption and a high abandonment rate.
Before the availability of OpenAI, the feature had limited functionality in understanding applicant input. The chatbot lacked conversational fluidity and often provided unhelpful responses, leading to poor applicant interactions. The chatbot setup for recruiters was fragmented across multiple pages, with a poorly designed information architecture that required users to check multiple locations. A lot of handholding from Customer Support Managers was needed to ensure the chatbot was activated successfully.
Urgency was further added to redesigning this Fountain AI feature when it was upsold to one of our biggest customers, who wanted to implement it within three months.
Outcome
My team rebranded the feature as Candidate AI Agent and I redesigned the chatbot setup flow for recruiters and the interaction flow for applicants, addressing specific user pain points identified through user interviews, competitive analysis of best practices for chatbot interactions, and a feature audit.
After an Alpha launch, closely collaborating with a single customer for initial feedback, we saw a 34% decrease in the recruiter’s manual SMS response rate and a 21% decrease in time-to-hire for applicants assisted by Candidate AI. This validated the project goal of reduced recruiter workload and improved applicant engagement. We are currently working towards our Beta launch with five more customers to refine the feature for a GA launch by October, 2024.
Team: 1 PM, 2 Engineers, 1 QA Engineer, 1 Designer | Role: Lead product designer | Platform: Mobile and Web application
Research & Findings
Customer interview (current + abandoned)
Audit of existing feature
Competitive analysis
In about two weeks in the research phase, I conducted feature audit, customer interviews and competitive analysis to understand what are the limitations with the existing Fountain AI feature that hindered customers to abandon it and the biggest pain points in its flow and functionality.
Design Goals + Process
Restructuring the information architecture of recruiter UX
Identifying ideal applicant journey
Iterating on design options
I collaborated with the PM and engineers to iterate on design options, focusing on enhancing UX for recruiters and applicants while addressing key challenges identified in the research phase.
Identifying and leveraging the new information architecture for recruiters and ideal applicant interaction flow, I developed and validated design options through prototype testing with internal stakeholders as the project advanced towards Alpha launch deadline.
Design Explorations
Final Designs
Alpha Launch Results
With the redesigned flows for recruiters and applicants, we Alpha launched the Candidate AI feature to a single customer for focused testing and feedback. This decision allowed us to collaborate closely with a customer to gather insights with an emphasis on product refinement before broader GA release.
A month into the launch, we observed a large volume of applicants interacting with Candidate AI from the customer’s career site and a 34% decrease in the recruiter’s manual SMS response rate and a 21% decrease in time-to-hire for applicants assisted by Candidate AI.
We are currently collecting feedback from our Alpha customer and developing towards Beta launch based on the feedback with five more customers.