AI-Powered Scorecard
Automated candidate assessments, allowing recruiters to prioritize critical tasks like building relationships with clients and talent.
Bolster
website
Bolster is an AI recruitment platform that accelerates the executive search process.
My Role
Design System
Discovery
Product Design
Project Management
Research
Team
Chief Technology Officer
Data Scientist
Lead Engineer
Product Manager
Timeline
3 Months
2024

AI-Powered Scorecard

We automated candidate assessments to save time and allow recruiters to prioritize critical tasks like building relationships with clients and candidates.

Outcomes

Optimizing the candidate assessment process reduced time to first candidate introduction and time to hire by 20%.

First Candidate Introduction
Reduced
20%
Time to Hire
Reduced
20%

The Process

As Bolster expanded its executive search business, we looked for ways to enhance the recruiting team's effectiveness. During user interviews with recruiters, we discovered that writing candidate assessments was one of the most time-consuming aspects of their roles.

Building Trust

For this project, it was essential for recruiters to trust the AI output from the scorecards. To address this, we created prototypes so recruiters could review the output, provide feedback, and enable us to iterate on it before developing any UI on the platform.

Airtable and Streamlit prototypes were used to test the scorecard concept and collect feedback on the output.

Ruling Out Assumptions

Creating prototypes allowed us to test an early concept that recruiters would want to rate candidates based on their CXO Persona type. In the testing phase, recruiters and clients didn't find the assessments valuable. Ruling this out in the prototype phase saved us valuable time later in the development phase.

Notes kept throughout the discovery, research, and testing phases.

Learnings

Data didn't give us a complete picture of a candidate's background. We learned that screening calls gave recruiters additional insights that would influence scores. We updated the prototype to incorporate candidate notes already stored on the platform, which provided us with more accurate candidate scoring.

Early wireframe concepts for scorecard scale.

Solution

We automated the candidate assessment process by creating an interface that allows recruiters to input freeform candidate requirements. Using these requirements, we leveraged OpenAI to analyze candidates' experiences and data about the companies they have worked for. This solution enables recruiters to quickly evaluate and score candidates while also eliminating any unintended bias. 

The final version of the productized candidate scorecard experience.
Candidate Scorecard Tab: processed, edit, and empty state.
Requirements are added at the project level and processed for individual candidates as needed.