Case Study:
FairNow & Humanly

About Humanly

Humanly, a trailblazer in conversational AI for hiring, elevates the recruitment process with its innovative AI-driven capabilities.

With features such as candidate screening, scheduling, and note-taking, Humanly is designed to automate the heavy lifting in the recruitment process, elevating both the candidate experience and recruiting team efficiency.

Context

Late 2023 and early 2024 saw a significant uptick in AI regulations and standards, especially with the designation of ’employment’ as a high-risk sector of AI applications. This development coincided with prominent lawsuits, such as Mobley v. Workday and Moffatt v. Air Canada, sparking increased oversight of AI-driven candidate screening and conversational AI.

The Challenge

Establishing Trusted AI in an Era Of Increased Scrutiny

Employers wanted assurance that their vendor technology was unbiased. Humanly identified an early advantage in demonstrating their product’s trustworthiness to their customers.

Humanly partnered with FairNow to tackle two primary goals: rigorously assess their existing AI governance protocols and conduct a comprehensive bias assessment of their AI-enabled conversational chatbot.

Guru Setupathy, the CEO of FairNow, points out, “What makes Humanly special is that they were not simply reacting to compliance requirements; they were dedicated to building trust in the candidate experience, which is a notable differentiator in the HR technology space.”

This strategic approach solidified Humanly as a leader in responsible HR tech practices and reinforced confidence in adopting their AI solutions, both across current customers and new sales opportunities.

“Just as humans ‘are what they eat, ‘ AI is only as good as its data quality, relevance, transparency, and explainability,” says Prem Kumar, CEO of Humanly.

“It is up to AI providers to keep pushing these ethical standards, grounding transparency and explainability in facts. This fosters human connection and trust while minimizing risk to deliver real-world outcomes.”

The Solution

Step 1: AI Governance Evaluation

FairNow’s collaboration with Humanly began with a detailed AI governance assessment of their conversational AI platform.

Our assessment analyzed and applied a vast array of laws, standards, and best practices relevant to their use case.

Our goal was straightforward: to pinpoint and address any possible compliance or governance gaps within Humanly’s operations.

Step 2: Chatbot Bias Evaluation

Secondly, FairNow conducted a bias evaluation of Humanly’s AI chatbot. The automated bias evaluation used FairNow’s proprietary IP to interact with Humanly’s AI chatbot using various questions and scenarios designed to probe for bias.

FairNow’s testing generated synthetic data specifically designed to catch bias based on how users interact with the Humanly chatbot. Humanly’s work with FairNow is helping set the standards to demonstrate the safety of AI using red-teaming techniques.

Results

Based on FairNow’s evaluation, Humanly’s generative AI governance is high-quality, rigorous, and low-risk as of Q1, 2024.

Humanly’s enthusiastic adoption of FairNow’s recommendations demonstrated their unwavering commitment to responsible AI development and constant refinement.

Bryan Leptich, CTO of Humanly, who was involved with the implementation, shares,

“Our ultimate goal is to be a thought leader in the application of ethical and responsible AI. This means going far beyond legal requirements in how we speak, act, and create. We believe our work with FairNow facilitates this very goal.”

The collaboration between Humanly and FairNow helped establish new benchmarks for responsible AI usage in recruitment.

Join industry-leading organizations like Humanly by demonstrating robust AI governance and compliance standards.