Pre-employment English test: what to look for

Resumes tell you someone studied English. A certificate tells you they passed a test on a specific day. Neither tells you whether they can run a customer call, handle a support ticket over the phone, or sit in a standup and be understood.

That gap is what a pre-employment English test is supposed to close. Most of them don't, because they test the wrong thing. Here is what to look for before you put one in front of a candidate.

What a good pre-employment English test measures

Start with the job. If the role involves talking to people, the test has to involve talking. A reading-and-clicking exam can confirm someone knows grammar rules, but it cannot tell you whether they freeze when a customer interrupts them.

Look for a test that scores spoken English across the skills that actually carry a conversation:

• Grammar and vocabulary, so you know the range they can express ideas in.
• Fluency and coherence, so you know they can keep a thought moving under pressure.
• Comprehension, so you know they follow what the other person is saying.
• Task completion, so you know they can actually get something done in the conversation.

Notice what is not on that list: accent. A useful test judges whether someone can be understood and can understand, not whether they sound like a particular region. We go deeper on the mechanics in how to assess spoken English.

Speaking vs. multiple-choice

Multiple-choice and gap-fill tests are popular because they are cheap to build and trivial to grade. They are also easy to study for and easy to game. Someone can ace a grammar quiz and still struggle to string two sentences together on a live call.

A speaking test flips that. The candidate has to produce real language in real time, which is exactly the signal you are hiring for. The trade-off used to be cost: live interviews take a recruiter's time, and they are hard to run consistently across dozens of applicants.

That is the part an AI voice interviewer solves. The candidate talks through about ten questions in 10 to 15 minutes, and the transcript is scored automatically. You get the signal of a real conversation without booking a slot on anyone's calendar.

Which level to require

Anchor the test to a standard instead of a gut feeling. CEFR is the common one: A1 to A2 is basic, B1 to B2 is independent, and C1 to C2 is proficient. Most hiring bars sit at B2, which roughly means the candidate can hold a fluent work conversation without strain.

Two mistakes are common. The first is demanding C1 for a job that runs perfectly well at B2, which shrinks the pool and raises salaries for no gain. The second is not stating a level at all, so every hiring manager applies their own. If the role is customer-facing, our guide to English levels for customer service jobs walks through the bar channel by channel.

Fairness and consistency

A test is only fair if every candidate gets the same one. Human screeners drift: the third interview of the day is graded differently from the first, and two recruiters rarely agree on what "good enough" sounds like. That inconsistency is a legal and quality risk, not just a nuisance.

Consistency also means you can defend a decision. When the output is a CEFR level, six skill scores, a short list of strengths and weaknesses, and the recording itself, you have a record you can point to, not an impression you have to justify from memory.

What the report should show

The test is only half of the product. The other half is what lands on your desk afterwards. A report that just says "72%" leaves you guessing what that means for the job. Look for four things:

• An overall CEFR level, so the result maps directly to the bar you set for the role.
• A score per skill, so you can see that a candidate is strong on comprehension but thin on vocabulary, or the reverse.
• Concrete strengths and weaknesses in plain language, written for a hiring manager rather than a linguist.
• The recording and transcript, so anyone on the team can listen to the candidate for two minutes and hear the level for themselves.

You can see a sample report to judge whether that level of detail is what your team needs.

Candidate experience

The best signal in the world is useless if good candidates abandon the test. Watch for friction. If a pre-employment English test forces an account signup, a software install, or a scheduled proctoring slot, you will lose strong applicants before they start.

Aim for a one-time link the candidate can open on any device, with no account required. Ten to fifteen minutes is a reasonable ask; an hour-long exam is not. The lighter the lift, the more representative your pool stays.

One thing to be clear-eyed about: the result goes to you, the recruiter, not the candidate. The report, the recording, and the recommendation are a hiring tool, so set expectations accordingly and keep the candidate-facing step short and respectful.

If you want to see what that looks like end to end, you can set up a screen and send yourself the link before you ever put it in front of an applicant.