Truffle calculates Match % using only job-relevant criteria you define. We never use demographic data, prestige signals, or appearance-based cues. Our goal is simple: surface your strongest candidates while reducing opportunities for bias to creep in.
We do this by being transparent about exactly what we analyze and exactly what we don’t.
What Truffle's AI analyzes
Here's what Truffle's AI analyzes.
Your job description
We extract key requirements, responsibilities, and qualifications you specify.
Success criteria questions
Your intake answers give us a clear picture of what makes someone successful in the role.
Candidate responses
Video or audio answers, along with any qualification responses, provide richer, role-relevant signals than a traditional résumé.
What Truffle's AI doesn’t analyze
We deliberately avoid using any information that could introduce bias; whether directly or through proxies.
Demographic data & protected traits
- Age or age indicators (e.g., graduation year, total years of experience beyond requirements)
- Sex, gender identity, sexual orientation, pregnancy or parental status
- Race or ethnicity, or proxies inferred from name, voice, or appearance
- Religion or religious affiliation
- Disability or health information
- Veteran status, political affiliation, or union membership
- Citizenship or immigration status (beyond legal work-authorization checks you require)
Appearance, environment & audiovisual cues
- Facial features, skin tone, hairstyle, tattoos, clothing, accessories
- Background environment or personal items visible in video
- Lighting, camera/mic quality, internet speed, or background noise
- Facial analysis, emotion detection, or micro-expression scoring
- Accent, pitch, tone, or speech rate beyond basic intelligibility
Socioeconomic & location proxies
- ZIP code, neighborhood, or country of origin
- Credit score or financial data
- Commute distance or car ownership
- Time-of-day submission patterns
Education & prestige factors
- University rankings or other “prestige” lists
- GPA cutoffs not tied to the job
- Brand prestige from bootcamps or courses
- School name unless you choose to display it (never scored)
Career history proxies
- Employment gaps as automatic negatives
- Senior-sounding titles without relevant skills
- Salary history or expectations (unless legally allowed and you opt in)
Online footprint & third-party data
- Social media accounts or follower counts
- Public records, background checks, or criminal history
- Web-scraped or purchased data
Other non-job-relevant signals
- Referrer identity or “network closeness”
- Writing style as a proxy for class or region (only scored if communication skills are a requirement)
- Personality or sentiment scores from AI tools
- Submission timing (weekday vs weekend)
Why it matters
Bias doesn’t just create unfair hiring; it creates poor hiring. Every irrelevant factor that enters the decision process increases the chance of missing a great fit. By removing these signals from analysis, Truffle keeps your shortlists consistent, explainable, and focused on what matters: the role and the candidate’s ability to succeed in it.
With Truffle, every Match % is grounded in the same evidence for every candidate: Clear, job-relevant criteria you set from the start.
