Skills-based hiring: how to implement trial tasks and workday simulations

71% of companies in Poland always or usually require a degree from specialists, and 85% complain about a skills gap. We show in six steps how to replace the “CV and degree” filter with a task that checks what a candidate can really do.

Who this guide is for and how long implementation will take

This guide is written for two organizational scales. The first is a small company without an HR department, where the owner or manager personally hires a few people a year. The second is a medium-sized or large organization with recruiters, line managers, and repeatable processes. Most steps are shared, and at the end each scale has its own version.

After reading, you will be able to review job ads for unnecessary requirements, describe a role as a set of competencies, design a trial task, assess it using fixed criteria, and check whether the change produced results. We also show where a fair recruitment task ends and unpaid work begins.

According to our estimate, a pilot for one position takes two to four weeks: one week for reviewing the job ad and creating the competency map, one week for the task and assessment criteria, and the rest for the first recruitment process. Measuring the results takes more time, because you will only know the performance of the hired person after the probation period.

The guide has the following sections:

1. Why degrees and CVs poorly predict job performance
2. Step 1: review job ads and remove unnecessary requirements
3. Step 2: the role competency map
4. Step 3: design of the trial task or simulation
5. Step 4: assessment criteria
6. Step 5: fairness toward the candidate
7. Step 6: measuring results
8. Version for a small company and for a large organization
9. Typical mistakes and next steps

Why degrees and CVs poorly predict job performance

Start with data, because you will need it in a conversation with the board or a business partner. The Hays Poland report “Labour Market 2026. Mid-year trend review” from July 2026, based on a survey among companies and specialists, shows a mismatch between requirements and reality. 71% of organizations always or in most cases require a university degree when recruiting for specialist and managerial positions. At the same time, 48% of specialists do not work in the profession they studied for, and 28% believe higher education is not necessary for success in their profession.

This filter has a cost. 30% of specialists have been rejected from recruitment at least once because they lacked the required education, even though they had the necessary competencies and experience. Another 21% do not know whether this happened, probably because they received no feedback. Meanwhile, 85% of companies report a skills gap, 14 percentage points more than at the beginning of 2026. Only 31% of companies say they are ready for recruitment based solely on competencies, and 44% cannot assess this.

The problem also lies on the university side. Prof. Ryszard Koziołek, rector of the University of Silesia, points out in a statement quoted by pulsHR.pl that from the decision to create a degree program to the first graduates entering the job market, six to seven years pass. A degree therefore confirms knowledge designed several years earlier, not the skills needed today.

The strongest argument comes from research on selection methods. The classic 1998 meta-analysis by Frank Schmidt and John Hunter compared predictive validity, that is, the strength of the relationship between a method’s score and later job performance (correlation coefficient r; the closer to 1, the stronger the relationship). Work sample tests reached r = 0.54, the highest among methods used individually. The structured interview scored 0.51, job experience 0.18, and education level only 0.10. The 2005 meta-analysis by Roth and colleagues confirmed the high validity of work sample tests. The 2022 reanalysis by Sackett and team lowered the absolute values for most methods, but trial tasks still rank near the top, ahead of the unstructured interview and references.

Company declarations should still be treated cautiously. The Harvard Business School and Burning Glass Institute report “Skills-Based Hiring: The Long Road from Pronouncements to Practice” from February 2024 showed that in the US in 2023 only about 1 in 700 hires resulted from dropping the degree requirement. About 45% of companies that announced a policy change changed it only on paper. The takeaway for you: simply removing the degree requirement from the job ad will do nothing if you do not change how candidates are assessed. That is what the next steps are for.

Step 1: review job ads and remove unnecessary formal requirements

Start with one position that you recruit for most often or find hardest to fill. Do not rebuild all job ads at once.

1. Open the current job ad and list each requirement in a separate row of a spreadsheet.
2. Next to each requirement, add what task in the role it is meant to safeguard.
3. Mark requirements that result from regulations, e.g. licenses in regulated professions such as medicine, law, or some engineering specializations. Do not touch those.
4. For the remaining formal requirements (degree, field of study, “minimum five years of experience”), ask: can a person without this, but with the skill, handle the task in the second column? If yes, remove the requirement or move it to the “nice to have” section.
5. Replace vague phrases with a description of tasks: instead of “stress resistance,” write “handling several urgent customer requests at the same time.”
6. Add to the job ad that the recruitment process includes a practical task, and state how long it takes.

AI

Screenshot: a spreadsheet with three columns - requirement, role task, decision (keep / remove / nice to have).

Expect that after this step the job ad will be shorter and more specific. Candidates with non-standard career paths will more often feel they can apply.

Step 2: build the role competency map

A competency map is a list of skills that determine performance in the role, along with a description of how each one appears in day-to-day work. Without it, the trial task will test random things.

1. Talk to two or three people who perform well in this role and to their manager.
2. Ask which situations in the last month had the greatest impact on their results.
3. From the answers, list four to six competencies. More than that is hard to assess reliably in a single process.
4. For each competency, add two or three observable behaviors, that is, behavioral indicators. Example for attention to detail: “checks the data in the system before replying to the customer.”
5. Mark which competencies the candidate must have from day one and which you will teach during onboarding.

This last distinction matters for the task design. Work sample tests are used mainly with people who already know the profession, and their validity has been studied best for them. If you are hiring a junior or someone from another industry, the task should test way of thinking and speed of learning rather than proficiency in a specific program.

Step 3: design a trial task or a workday simulation

The task should recreate a real situation from the role, not test textbook knowledge. Choose a format that matches the competencies from the map. The most commonly used types are:

• writing a specific text, e.g. a response to a complaint or an email to a contractor,
• data analysis with a decision recommendation,
• an in-basket exercise, that is, a set of matters to organize and handle within a limited time,
• role play, e.g. a conversation with an unhappy customer or a negotiation,
• a case study, that is, a description of a company problem with a request for a solution,
• a workday simulation combining several of the above.

Examples from industry practice: a candidate for a UX designer role (user experience design) evaluates the usability of three app screens, a sourcing candidate (active candidate search) selects LinkedIn profiles according to given criteria, and a manager analyzes team results and proposes a bonus distribution. In customer service, the Polish company Axendi describes on its blog group recruitment meetings with short tasks based on real situations.

Design the task in this order:

1. Select from the map the competencies the task should assess.
2. Describe the starting situation as it would look in your company, but using fictional data.
3. Set a time limit. For an unpaid task, aim for 30 to 60 minutes.
4. Prepare instructions for the candidate: what they should do, how much time they have, what materials they may use, including whether they are allowed to use AI tools.
5. Test the task on a current employee and measure how long it took them.
6. Shorten or simplify the task if the employee did not finish on time.

AI

You can build a workday simulation yourself or buy a ready-made one. One provider is the Polish start-up Talent Quest. According to the description in sponsored content on mamstartup.pl, the candidate receives a link to a browser-based dashboard with an email inbox, communicator, phone, and ERP and CRM systems (for managing company resources and customer relationships). The simulation lasts 30 to 60 minutes, and the company receives a report with a percentage score and suggested interview questions. For candidates, the tool is free of charge; the company pays for the full report on the selected person. This is an example of one provider, not independent proof of effectiveness. Before buying such a tool, check whether its tasks match your competency map and how the provider calculates the score.

Screenshot: a document with task instructions for the candidate - situation description, time limit, list of permitted materials.

Step 4: set the assessment criteria before you see the first solution

Trial tasks are easy to undermine with gut-feel assessment. The assessment center method, that is, the structured evaluation of candidates across several exercises, has developed rules that you can transfer to a smaller scale.

1. For each competency in the task, assign the behavioral indicators from the map.
2. For each indicator, describe levels, e.g. from 1 to 4, with an example of behavior at each level.
3. Assess the candidate against the role profile, not against other candidates.
4. Record what the candidate did before giving a rating.
5. If you can, assess each competency in two different tasks and involve two assessors who compare ratings only after scoring them.
6. Set the threshold for moving a candidate forward before the recruitment process starts.

Screenshot: an assessment sheet - competencies in rows, levels 1-4 in columns, a field for observation notes.

The task result does not replace the interview. Use it as material for a structured interview: ask the candidate why they made a given decision in the task. In the research by Schmidt and Hunter, combining several valid methods predicted performance better than any one of them alone.

Step 5: ensure fairness toward the candidate

A trial task takes the candidate’s time. If you do not respect that, the best candidates, who have several offers, may be the first to drop out.

Polish labor law does not provide for a “free trial day.” If a candidate is actually performing regular duties, e.g. serving customers or fulfilling orders, without a contract and pay, this is unlawful. If you want to test someone in real work, sign an employment contract for a probation period. Short recruitment tasks using fictional data are not precisely regulated, but industry articles agree on the general boundaries: 30 to 60 minutes is widely accepted as an unpaid stage, while more than two to three hours of work without pay raises doubts.

Apply these rules to every task:

1. State the task duration already in the job ad or in the first contact.
2. Use fictional data, not real customer assignments.
3. Offer payment for tasks longer than two hours.
4. Give the candidate a choice of timing, unless the time limit is part of the test.
5. Send everyone feedback with the result and one specific comment.

Feedback takes a few minutes per candidate. Hays data shows that 21% of specialists cannot even assess whether they were rejected from recruitment because of the lack of a degree.

Step 6: measure results and improve the process

Without measurement, you cannot distinguish real change from change on paper. Before the pilot, record the starting point from recent recruitment processes for the same position.

1. Count how many candidates received the task and how many completed it. A low completion rate usually means the task is too long or the instructions are unclear.
2. Measure the time from publishing the ad to signing the contract.
3. Check how many hires passed the probation period.
4. After three and six months, ask the manager to assess the new person on the same scale as in the task.
5. Compare the task score with the assessment after a few months. If there is no relationship, the task is testing the wrong things.
6. After a year, check retention, that is, how many hired people are still working.

AI

Retention is a good long-term metric. In the cited Harvard Business School and Burning Glass Institute report, in companies that had truly implemented skills-based hiring, workers without a degree had a retention rate 10 percentage points higher than people with degrees in the same roles.

Screenshot: measurement table - position, number of candidates, task completion rate, recruitment time, assessment after three months.

Version for a small company without an HR department

In a small company, your time is the biggest cost, so simplify every step. Build the competency map yourself in an hour, based on the work of a person who performs well in that role, or on your own work if you have been doing it yourself so far. Limit yourself to four competencies.

Prepare one task lasting 30 to 45 minutes, which the candidate completes remotely or during a meeting. For a sales representative, this might be replying to an inquiry for quotation; for an office assistant, organizing an inbox with a dozen or so messages. Assess using a simple sheet. If you do not have a second person for assessment, ask a business partner or a trusted employee for an independent assessment of two or three solutions.

Small companies are already more likely to look at practice. According to the EFL report “Education or experience? What helps in business. Under the magnifying glass,” 47% of entrepreneurs in the SME sector in 2025 attached greater importance to candidates’ practical skills than to theoretical knowledge, although 51% of companies still based decisions mainly on knowledge. Your advantage is speed: you can change the job ad and the task in a week, without approval from several departments.

Ready-made simulation platforms can save design time, but if you recruit only a few times a year, check whether paying for a report makes sense compared with a task you prepare yourself.

Version for a large organization

In a large company, the hardest part is keeping the process consistent across teams. Start with a pilot for one high-volume role, e.g. in customer service, sales, or logistics. Appoint a process owner, usually in the HR team, and a line manager responsible for the competency map.

Introduce a standard: a shared competency map template, a task library, an assessment sheet, and training for assessors. The rule of two assessors and two tasks per competency is realistic here. Agree with the legal department on the rules for payment for longer tasks and consent to use the results of the work.

The biggest risk is change only on paper. If recruiters still reject CVs based on the name of the university or previous employer, removing the degree requirement from the job ad will change nothing. That is why in the applicant tracking system you should move the practical task to an early stage, before CV screening or in parallel with it, and report the metrics from step 6 every quarter.

Typical mistakes

The same mistakes recur regardless of company size:

• the task tests theoretical knowledge instead of activities performed in the role,
• the assessment criteria are created after seeing the solutions, so they adapt to the favorite candidate,
• the task takes several hours and no one pays for it,
• the candidate is given a real customer assignment to do,
• the degree disappears from the job ad but remains in the mind of the person screening CVs,
• the company does not measure results, so it does not know whether the new method works better,
• candidates do not receive feedback.

Next steps

Choose one position today and list the requirements from its job ad in the spreadsheet from step 1. Within a week, talk to two people who perform well in that role and build the competency map. In the next recruitment process, run the first trial task, and after three months compare the task result with the manager’s assessment.

If the pilot works, extend the method to more roles and build a task library. If not, go back to the competency map and check whether it describes what performance in the role really depends on.