Now in private beta

Hire for AI fluency,not memory.

Give candidates real work tasks. Let them use AI. Pactum records every prompt and decision, then scores the session automatically.

Get started free

Live candidate session

Pactum - Assessment
01EasyAlgorithms

Two Sum

Given an array of integers and a target, return the indices of the two numbers that add up to the target. Each input has exactly one solution.

Constraints
You may not use the same element twice
Return indices in any order
Aim for O(n) time complexity
solution.js
JavaScript
1
2
3
4
5
6
7
8
9
10
11
12
13
14
AI AnalysisStrong hire

How it works

Three steps. Zero ambiguity.

Step 01

Set up in under 5 minutes

Pick task types that match the actual job: email, written analysis, code review, or slide deck. Add a prompt and tell Pactum what good looks like.

No test banks to license. No rubrics to write from scratch. Send your first assessment before your next meeting.

Pactum - Builder
Round 1: Task type
Email taskWritten reportCode reviewSlide deck
Task prompt

Success criteria
Tone matches the relationship stage
No pressure tactics or urgency framing
Clear next step with low friction
Step 02

See how candidates actually think

Candidates complete real tasks using built-in AI tools. Pactum captures every prompt, every revision, and every decision as they work.

You stop seeing only the polished final answer. You see the reasoning, the shortcuts, and the judgment calls behind it.

Pactum - Candidate workspace
Code review task

Review this SQL query for performance issues. The table has 40M rows and the query takes 12s.

AI AnalysisStrong hire
Technical depth92
AI Fluency89
Clarity87
Step 03

Make faster, more confident calls

Every session is automatically scored on prompt quality, output quality, critical thinking, and AI fluency. Candidates are ranked and ready to act on.

Stop debating cover letters. Compare actual work output from people who used the same tools your team uses every day.

Pactum - Results
Alex T.Senior PM
Strong hire
Communication94
AI Fluency91
Judgment88
Jamie S.PM
Lean yes
Communication78
AI Fluency72
Judgment80

The problem

Your hiring process was designed before AI existed.

LeetCode tests memory. Take-homes test endurance. Neither one tells you whether a candidate can do the job now that AI is part of every job.

Old way

Pactum way

Tests memorization and endurance
Tests real work output with AI tools
You see the answer, nothing else
You see every prompt, revision, and decision
Manual review takes hours per candidate
Every session scored automatically in seconds
Top candidates skip long unpaid take-homes
Async, focused tasks candidates actually respect

The research

AI does not replace workers. It makes skilled ones dramatically better.

Independent randomized controlled trials confirm large, replicable gains in speed and quality. The catch: the gains go to workers who know how to use AI well.

0%

faster on writing tasks

In a preregistered RCT with 453 college-educated professionals, ChatGPT access cut task completion time by 40% and raised output quality by 18%, graded by independent human evaluators.

Science, 2023

Noy & Zhang, MIT Department of Economics

>0%

higher quality scores

758 BCG consultants using GPT-4 completed work rated over 40% higher quality. Below-average performers improved 43%. Tasks outside AI's capability range saw a 19-point drop (the "jagged frontier").

Organization Science, 2026

Dell'Acqua et al., Harvard Business School / BCG

0%

more issues resolved per hour

Across 5,172 customer-support agents, AI access raised hourly productivity 14% on average, with a 34% gain for novice workers. Customer sentiment scores improved alongside throughput.

Quarterly Journal of Economics, 2025

Brynjolfsson, Li & Raymond, Stanford / MIT

0%

wage premium for AI-skilled workers

Analysis of over 1 billion job postings found AI-required roles command a 56% wage premium, up from 25% the prior year. AI-skilled jobs grew 7.5% even as total postings fell 11%.

Global AI Jobs Barometer, 2025

PwC, analysis across 6 continents

Studies used GPT-3.5 through GPT-4 era models. Directional findings (that skilled AI use outperforms unskilled AI use) are consistent across model generations and independently replicated.

Pricing

Simple, transparent pricing.

Start free. Scale when you're ready.

Starter

$99/month

For teams hiring 1–5 roles per month

Up to 5 active assessments
All workspace types
Unlimited candidates
AI scoring & rubrics
Email support
Most popular

Professional

$549/month

For teams scaling hiring across departments

Unlimited assessments
All workspace types
Unlimited candidates
AI scoring & rubrics
Custom rubric builder
Priority support

FAQ

Common questions.

Everything you need to know before your first assessment.

Early access

Be first.
Hire better.

We're onboarding a small number of companies in our private beta. Spots are limited.