Skip to main content
VirtualInterview.ai

Software & Engineering

IT support and help desk mock interview

An IT support or help desk interview tests how you troubleshoot out loud: the questions you ask first, the order you check things in, when you escalate and how you explain the fix to someone who is not technical. Expect a phone screen, then spoken scenarios such as a slow PC, a VPN that will not connect, a locked account or a phishing report.

The practice link on this page opens a technical round with the role set to IT Support Specialist. In that round each main question is a fault to diagnose or a scenario to reason through, with at most one “what is X?” warm-up, and the interviewer follows up when an answer skips a step. Each answer is scored from 0 to 100 for relevance, depth and clarity. When no code is written, the round report weighs problem solving (40%), technical knowledge (33%) and communication (27%); code quality only counts if you write code, which a help desk round rarely asks for.

Practice this role with an AI interviewer

Free plan: 2 interviews a month, free account, no card.

What the loop usually looks like

  1. Phone screen (Short). Your experience and certifications, the shifts or on-call you can cover, and how you talk to users. Some screens slip in one quick scenario, such as a user who cannot log in.
  2. Technical or troubleshooting interview (Often the longest round). Spoken scenarios: talk through a fault, say what you would check and in what order, and what you would do when a check comes back clean. Some employers add a hands-on task or a written ticket exercise.
  3. Team or manager interview (Varies). Ticket handling, a frustrated user, documentation, how you escalate to second line or another team, and what you do when a senior person needs something fixed in five minutes.

What interviewers score

  • A structured troubleshooting order, said out loud
  • Clarifying questions before the first fix
  • Explaining the fix to a user who is not technical
  • Tickets, documentation and knowing when to escalate

What interviewers probe at each level

Tier 1 help desk

Interviewers check method more than knowledge. Expect a user who cannot log in, a printer that will not print and a slow laptop, and they listen for whether your first move is a question (“is it just you, or others too?”) or a fix. CompTIA A+ is the usual entry certification, and CompTIA says it appears in more tech support job listings than any other IT credential. If you are studying for it, say which of its two exams you have passed.

Desktop support or second line

The questions move to root cause and tools: imaging and deploying machines, Active Directory and group policy, an account that locks every morning, a driver that breaks after an update, hardware swaps. Expect “the first-line fix did not work, now what?” and a question on how you write up a fix so first line can use it next time.

Service desk lead

Lead rounds ask how you run the queue: ranking tickets by impact and urgency, keeping to response targets, building the knowledge base, and spotting the same incident across many tickets so a problem record gets opened. Bring one recurring issue you traced and what fixing it did to ticket volume.

Mistakes that cost candidates this round

  • Jumping to a fix (“I would reset the password”) before asking a single question.
  • Listing checks in no order, so the interviewer cannot tell what you would do first.
  • Saying “I would escalate it” without saying what you already tried and wrote down.
  • Unlocking the same account again and again without finding what keeps locking it.
  • Explaining a fix to the user in jargon.
  • Treating a phishing report as a nuisance instead of asking whether the user clicked or typed a password.

How do you answer help desk troubleshooting questions out loud?

Use the same order every time and say it as you go. Start by asking what exactly happens, since when, what changed and whether anyone else has it. Then name the most likely cause and the cheapest way to test it. If the test rules it out, move to the next cause. When something works, confirm with the user that everything works, not only the symptom they called about, and write the ticket up. Interviewers listen for the next check: say “I would restart it” and the follow-up is “it is still slow after a restart, now what?” The five walkthroughs below give the order and the next check at each step.

For a slow PC, ask whether everything is slow or one program, since when, and whether anything changed, such as an update or new software. Open Task Manager and see what is using processor, memory or disk. If one process is pinned, deal with that process. If nothing stands out, check startup programs, pending updates and free disk space, then run a malware scan. Still slow after that, check disk health and the hardware, and decide whether reimaging is quicker than searching further. Tell the user what you found in one sentence.

For a VPN that will not connect, first check whether the internet works without the VPN. If it does not, you have a network problem and the VPN is a symptom. If it does, read the exact error message, confirm the username, password and multi-factor prompt, and check the client version and the device clock, since a wrong date can break certificate checks. Then ask whether other users have the same problem. If they do, stop working on one laptop and raise it as a wider incident.

For a locked account, verify who you are talking to before you unlock anything. Unlock or reset as your policy allows, then ask whether the password changed recently. Thresholds are set per organisation. Microsoft’s documentation for Entra Domain Services, for example, gives its defaults as a lockout after five bad passwords in two minutes, with an automatic unlock after 30 minutes, and names the usual causes: a user retrying an old password, or an application or service still using it. A phone syncing mail with the old password is a common culprit. If the account keeps locking, trace the source from the lockout event (ID 4740) rather than unlocking it again.

For a printer, ask whether it is one user or everyone. Everyone points at the printer or the print server: check that it is on and online, has paper and shows no error, and look for a stuck job in the queue. One user points at their machine: the default printer, the driver, a stuck local queue, and restarting the print spooler service. For a network printer, check that it still has the address the computers expect.

For a phishing report, thank the user first, because you want them to report the next one too. Ask whether they clicked a link, opened an attachment or typed a password. If they did none of those, have them use the report button, delete the message, and check whether colleagues received it. If they entered a password, reset it, sign them out of active sessions and pass it to security straight away. CISA lists the signs to point out: urgent or emotional wording, requests for personal or financial information, and addresses or links that are slightly wrong. It adds that perfect grammar no longer rules a message out.

Expect one quick definition question as well, often about networking. Answer it in a sentence, then tie it to a fault, because that turns a quiz answer into troubleshooting. DNS turns names into addresses, so if a user can reach a site by its IP address but not by its name, suspect DNS. DHCP hands out addresses, so a Windows machine showing a 169.254 address in ipconfig /all did not get one, and you check the cable, the Wi-Fi connection or the DHCP server before anything else. Ping tells you whether a device answers at all, which splits “the network is down” from “one application is down”.

Sources: Microsoft Learn, Troubleshoot account lockout problems in Microsoft Entra Domain Services; CISA, Recognize and report phishing

How do you prioritise tickets, and what separates an incident from a problem?

A common question: three tickets arrive at once. A department cannot run payroll, a new starter has no laptop on their first morning, and an executive’s mouse has stopped working. Which first? A common way to decide is to rank by impact (how many people, how critical the service) and urgency (how fast the harm grows). Payroll is high on both, so it goes first. The new starter is next because the morning is passing. The mouse is low impact even for an executive, but it takes a minute, so a colleague can drop off a spare. Say the rule, apply it, and add that your desk’s own priority matrix and service targets decide in the end.

ITIL 4 defines an incident as an unplanned interruption to a service, or a reduction in its quality, and a problem as a cause, or potential cause, of one or more incidents. Incident management restores the service quickly; problem management looks for the underlying cause so the incidents stop recurring. Give an example in the interview. Ten users who cannot print are incidents, and restarting the print server resolves them. The print server running out of memory every Monday morning is the problem.

Sources: Global Knowledge, How ITIL differentiates problems and incidents

How do you handle an angry user on the phone?

Many help desk loops include one: a user who lost an afternoon’s work, or a manager who missed a client meeting because the laptop would not start. Let them finish, name what it cost them, own the next step, explain the cause in plain words once you know it, and give a time for the next update. The US Bureau of Labor Statistics lists patience with frustrated users, and explaining fixes to people who are not technical, among the qualities the job needs.

To rehearse that call, use the angry customer role-play. Set the issue to something from your world, such as a laptop that would not boot before a presentation. The AI customer pushes back on vague answers and wants a time for the fix, and the report scores empathy (25%), ownership (25%), clarity (20%) and commitment (30%).

Sources: US Bureau of Labor Statistics, Computer support specialists; CompTIA, A+ certification

10 it support specialist interview questions and model answers

1. A user calls and says they cannot log in. What do you do first?

Technical. What they are testing: Clarifying questions before the first fix.

I ask before I touch anything. What exactly happens: an error, a wrong-password message, or the screen just hangs? Is it this computer only, or everything, including email on their phone? Did they change their password recently? Then I verify who they are. If the account is locked, I unlock it as policy allows and check where the lockouts are coming from. If the password has expired, I walk them through the reset. If the account is fine and the computer is the problem, I check its network connection and whether it can reach the domain.

2. Walk me through a PC that has become very slow.

Technical. What they are testing: A troubleshooting order with the next check after each step.

First, is everything slow or one program, and since when? Did anything change, like an update or new software? Then Task Manager: if one process is using all the processor or memory, I deal with that. If nothing stands out, I look at startup programs, pending updates and free disk space, and run a malware scan. If it is still slow, I check the disk’s health and the hardware. At some point reimaging is faster than hunting, and I would say so to the user and agree a time. Whatever it was goes in the ticket.

3. A remote user’s VPN will not connect. Talk me through it.

Technical. What they are testing: Separating a local fault from a wider outage.

I check whether their internet works without the VPN. If not, it is their connection, and I help with that first. If it works, I ask for the exact error text, confirm their credentials and that the multi-factor prompt arrives, and check the client version and the laptop’s date and time. Then I check whether anyone else is reporting the same thing. If several people are, it is probably the VPN service, so I raise it as an incident and tell the user what I am doing instead of reinstalling their client.

4. The same user’s account locks every morning. What is going on?

Technical. What they are testing: Finding the cause instead of repeating the fix.

Something is still trying an old password. Usually it is a phone or tablet syncing email, a mapped drive or saved credentials on another computer, or a service or scheduled task running under their account. I ask when they last changed the password and which devices they use, then look at the lockout event, ID 4740, for the source. Once I find the device, I update or remove the old password there. Unlocking every morning only hides it.

5. A user forwards an email they think is phishing and says they may have clicked the link.

Situational. What they are testing: Calm, quick containment and a user who will report again.

I thank them for reporting it and ask exactly what they did: clicked only, opened an attachment, or typed a password on the page it opened. If they entered a password, I reset it, sign them out of active sessions and pass it to security straight away, with the original message attached as our process asks. If they only clicked, I still tell security and check the machine as instructed. Then I check whether others received the same email so it can be pulled from their mailboxes.

6. Three tickets arrive at once: payroll is down for one department, a new starter has no laptop, and an executive’s mouse has stopped working. Which first?

Situational. What they are testing: Prioritising by impact and urgency, and saying the rule.

Payroll first. It stops a whole department and the harm grows by the hour. The new starter next, because their first morning is going. The mouse is low impact even for an executive, but it is a one-minute fix, so I would ask a colleague to drop off a spare while I work on payroll. I would follow our priority matrix if it says otherwise, and I would update each person on when they will hear from us.

7. What is the difference between an incident and a problem?

Domain. What they are testing: ITIL 4 vocabulary used correctly, with an example.

In ITIL 4 an incident is an unplanned interruption to a service or a drop in its quality, and the goal is to get the service back quickly. A problem is the cause, or possible cause, of one or more incidents, and the goal is to find it so the incidents stop. If a printer jams every Monday, each Monday’s ticket is an incident that I resolve. Noticing the pattern and getting the cause fixed is problem management.

8. You cannot fix an issue and nobody senior is around. What do you do?

Situational. What they are testing: Honesty, a workaround and a ticket the next person can use.

I tell the user I have not fixed it yet, what I have tried and when they will hear from me. If there is a workaround, such as a loan laptop or webmail, I set it up so they can keep working. I write the ticket so the next person can start where I stopped: the symptoms, each check and its result, and what I suspect. Then I escalate the way the process says and follow up at the time I gave.

9. An executive cannot join a video call that starts in five minutes.

Situational. What they are testing: Workaround first, diagnosis later.

I get them on the call any way that works: the browser version, the phone dial-in or a spare laptop. Once they are in, I log a ticket and look at the original fault after the meeting, when there is time to check the audio device, the app version or the permissions properly. Five minutes is not the time to start reinstalling software.

10. How do you explain a technical fix to someone who is not technical?

Behavioral. What they are testing: Plain language, and checking the user can carry on alone.

I describe what they will notice, not what I changed. Instead of “I cleared your cached credentials”, I say “your computer was still trying your old password, so I removed it, and it will ask for the new one once.” I check they can do the next step themselves, and if it might happen again I tell them the one sign to watch for. Then I put the same plain explanation in the ticket so it can go into the knowledge base.

Frequently asked questions

What questions are asked in a help desk interview?

Mostly troubleshooting scenarios you answer out loud: a user who cannot log in, a slow PC, a VPN or printer fault, a phishing report. Add a ticket priority question, incident versus problem, a frustrated user and what you do when you cannot fix something. The ten questions on this page cover each with a model answer.

Do I need a certification for an IT support job?

Not always. The US Bureau of Labor Statistics says some computer user support jobs accept a high school diploma with relevant certifications or experience, while others ask for a degree. CompTIA A+ is the common entry-level choice, and CompTIA says it appears in more tech support job listings than any other IT credential.

How is a desktop support interview different from a help desk interview?

Help desk interviews focus on remote troubleshooting and talking to users. Desktop support, often second line, goes further into hardware, imaging, Active Directory and group policy, and finding the root cause of something first line could not fix. Expect “the first fix did not work, now what?” more often.

Can I practise help desk scenarios with AI?

Yes. The link on this page opens a technical round for an IT support specialist, where the questions are faults and scenarios to reason through and the interviewer follows up when you skip a step. You can answer by voice or by typing. Free plan: 2 interviews a month, free account, no card.