Mentl gives teams two things: a standardized way to assess candidates on AI ability, and a training platform to upskill employees already on your team.
The problem
"Proficient in AI tools" tells you nothing about whether a candidate can write a prompt that actually does the job.
Asking "how do you use AI in your work?" produces rehearsed answers — not demonstrated ability under realistic conditions.
A Mentl score is tied to domain and difficulty. Compare candidates across your entire applicant pool with confidence.
How it works
Choose from domain-specific challenge sets — Finance, Legal, Engineering, Marketing, and more. Each challenge tests both subject-matter expertise and ability to apply it through AI.
Candidates receive a timed challenge link. They write a prompt that demonstrates domain knowledge and effective AI use. The grading engine scores it against a professional reference answer.
Every submission scores across rubric accuracy, prompt efficiency, and completion time. You get a breakdown that is consistent, comparable, and defensible — not a gut feeling.
The framework
Every challenge is designed to assess two things simultaneously. A candidate cannot fake either one.
Each challenge embeds real domain context — financial models, contract clauses, system logs, campaign data. Candidates who do not know the subject cannot write an effective prompt. There is nowhere to hide.
Knowing the domain is not enough. The grading engine scores how precisely candidates translate that knowledge into a prompt — specificity, structure, constraint clarity, and output framing all matter.
Pricing
Hiring assessments
$3 / credit
1 credit = 1 assessment invite. Credits never expire. Pay only for what you use.
Growth Pack 25 credits · Scale Pack 100 credits
Employee training
$5 / seat / month
Scales with your team. 10 seats = $50/mo · 50 seats = $250/mo · 200 seats = $1,000/mo.
Includes AI-generated programs, certifications, and analytics
Domains
Sign in, create your organization, and you'll have 5 free assessment credits waiting.