The AI capability assessment platform that scores real work

With Codility AI for Business, people in marketing, sales, finance and operations complete work simulations using AI, and their output is scored against defined test cases.

Simulation report Marketing operations
82 / 100
check_circle Brief reframed into a workable task
check_circle Output matches the required format
check_circle Source figures carried through correctly
cancel Edge case in the exclusion rule missed
Test cases passed 9 / 11
AI interactions
00:41 Asked the model to summarise the meeting notes
03:12 Constrained the output to the reporting template
07:55 Checked the figures against the source file
A business-role simulation report. The score comes from the work produced and the test cases it has to pass, with the AI interactions there for review.

Trusted by teams at GitHub, SpaceX, Tesla, EY, SAP, Barclays, and Citi.

Trusted by GitHub, SpaceX, Tesla, EY, SAP, Barclays, Citi.

G2 Leader, Enterprise
G2 Most Implementable, Enterprise
G2 Best Estimated ROI, Enterprise
G2 Fastest Implementation, Enterprise

Codility is rated 4.6 of 5 across 865+ G2 reviews and ranked #1 for Enterprise Technical Skills Screening. Those rankings come from verified customer reviews, not paid placements, and G2 updates them quarterly.

The board asked who can actually work with AI. The dashboard can’t answer.

Your talent team has four ways to answer the board today, and each one measures a proxy.

01

Training completion dashboards

The licenses are on and the courses are done. Completion proves attendance. Whether anyone can apply what they clicked through is a question the dashboard can’t answer.

02

Multiple-choice AI quizzes

Candidates answer ten minutes of definitions and tool trivia. A quiz score tells you who studied.

03

Self-report questionnaires

Self-assessments show how confident people feel about using AI. A hiring or promotion decision also needs evidence of what they can do.

04

Unstructured AI interview questions

When managers ask different questions about AI skills and use no shared criteria, their assessments are difficult to compare.

None of these four ways answers the question the board asked: who here can get work done with AI.

Give people real work, with AI in the loop, and score what comes out

An AI capability assessment platform measures whether specific people, candidates and employees, can genuinely get work done with AI. Organizational AI maturity asks a different question, about company strategy and data. A knowledge check asks a third one, about whether someone can define the terms.

AI for Business gives you verified evidence of who can get real work done with AI, in every role that uses AI.

People complete real work simulations with AI available. The score comes from the work they produce, measured against test cases the same way every time. Every AI interaction is captured as reviewable activity.

The score is evidence you can look at, because the score is the work product itself. Codility has run this assessment engine at enterprise scale since 2009. AI for Business is live today.

Test whether candidates can deliver work with AI, before the offer

  • Add one assessment to the hiring process for any business role.
  • Candidates work through a simulation with AI available, the way the job actually works now: they frame the task, work with the AI, and submit real output.
  • The assessment plugs into the applicant tracking system you already run, including Greenhouse, Lever, Ashby, Workday, and SmartRecruiters.

You stop hiring on a CV claim of “proficient with AI.” The shortlist is built on work someone actually produced.

Hiring pipeline REQ-4471
Marketing Manager 4 stages
1 Application review
2 AI work simulation Added
3 Hiring manager interview
4 Offer
check_circle Synced to your applicant tracking system
One added step in the hiring process: a work simulation with AI in the loop.
Task library AI for Business · 2 selected
check_box AI meeting-tool task
Turn a raw meeting transcript into a decision summary.
Role-agnostic 30 min
check_box Compliance-prompt task
Draft a policy-safe customer reply from a source document.
Role-agnostic 25 min
Marketing · Sales · Finance · Operations
Role-agnostic tasks: the same problem-solving standard across business functions.

Test problem solving with AI, without building a task per department

  • The AI for Business tasks are role-agnostic by design: an AI meeting-tool task, a compliance-prompt task, the kind of knowledge work that turns up across business functions.
  • What they measure is whether someone can frame a problem, work with AI on it, and produce something usable.
  • Common tasks let you assess problem solving with AI across marketing, sales, finance, and operations.

One yardstick covers marketing, sales, finance, and operations. You commission nothing before you start assessing.

See the evidence behind each score

  • Scoring is deterministic: work is measured against test cases, so the same submission produces the same score every time, with no generative model deciding who passes.
  • The AI interactions from the assessment sit alongside the score, so a reviewer can see how someone got there.
  • Codility scores the work; you make the decision.
  • The platform doesn’t rank people out, build the shortlist, or decide on a hire.
  • Codility holds SOC 2 Type II and ISO 27001 and operates under GDPR. You choose EU or US data hosting, with EU data held in Frankfurt. Codility trains no AI models on your candidate or employee data.

The record of the decision holds up when a candidate, an auditor, or your own legal team questions it.

Score breakdown 9 / 11 passed
TC-01 check_circle Required fields present in the output
TC-04 check_circle Figures match the source file
TC-07 cancel Exclusion rule applied to edge case
TC-11 check_circle Output format matches the template
AI interactions
02:18 Asked for a first draft
06:03 Narrowed the scope
11:47 Verified the numbers
What produced the score: the work and the test cases it passed. The AI interactions are there for review.

How this differs from quiz and questionnaire tools

01 The score comes from work someone actually produced

  • Quiz tools score recall.
  • Questionnaire tools score self-description.
  • Buyers report that some tools in this category have a generative model grade the output, so the same submission can come back with a different score on a second run.
  • Codility measures real work product against test cases.
  • A score that changes between two runs of the same submission is a score you cannot defend.

02 A new category, on a proven engine

  • Buyers report that most of the AI skills assessment tools on their shortlist shipped this year.
  • AI for Business runs on the assessment engine Codility has operated at enterprise scale since 2009, and scores through the same deterministic test-case method Codility customers already use.
  • You’re not trusting a hiring decision to a brand-new system.

What Codility customers say

Don’t test whether my people know the AI tools. Test whether they can actually solve a problem with them.

Enterprise customer, on what an AI capability assessment has to do

Codility is rated 4.6 of 5 across 865+ G2 reviews and ranked #1 for Enterprise Technical Skills Screening. Same platform, same scoring engine, now assessing AI capability in every role that uses AI.

Frequently asked questions

What is an AI capability assessment platform?

An AI capability assessment platform tests whether specific people can get real work done with AI, and produces a score per person. People complete role-relevant work with AI available, and the output is evaluated. The result tells a hiring team which candidates can deliver work with AI for hiring decisions and workforce baselines.

Run your first AI work simulation

Tell us which business roles you want to assess. A Codility specialist sets up your trial and loads the AI for Business tasks within one business day.

Start your free trial

A Codility specialist will be in touch within one business day to activate your trial.