| License required | None. No state or national license exists for help desk work. The de facto credential is CompTIA A+, and most postings word it as "A+ or equivalent experience" rather than as a hard requirement. |
|---|---|
| CompTIA A+, what it takes | Two separate exams, Core 1 and Core 2. You hold the certification only after passing both, and each attempt costs a few hundred US dollars at list price. CompTIA renumbers and retires the exam pair roughly every three years, so check the current exam codes and download the objectives PDF from CompTIA's own site before you buy any course or practice test. |
| Time to get credentialed | Two to four months of part-time study for both A+ exams is realistic for someone starting from ordinary consumer computer confidence, faster if you already repair your own machines. Add about a month of evenings to build and write up the home lab that makes the certificate mean something in an interview. |
| Degree requirement | Not standard. The US Bureau of Labor Statistics does not list a bachelor's degree as the typical entry-level education for computer support specialists, and OPM's qualification standard for the federal GS-2210 IT series asks for IT-related experience demonstrating four competencies: attention to detail, customer service, oral communication and problem solving. Some large enterprises and universities still filter on a degree in their HR system; MSPs, staffing firms and small-company IT generally do not. |
| Pay, where to get a real number | US BLS Occupational Employment and Wage Statistics, SOC code 15-1232, Computer User Support Specialists. Note that 15-1231 is Computer Network Support Specialists, a different and generally better-paid code. OES publishes median and percentile wages by state and metro area annually. For a live local range, filter job boards to pay-transparency states such as Colorado, California, Washington, New York, Illinois and Minnesota, where the ad must carry a range; the list of those states keeps growing, so check your own. |
| How the work is structured | Usually hourly and non-exempt, so overtime is paid. Tiered as T1/T2/T3 with a documented escalation path, and measured: first-contact resolution, time to resolve, reopen rate, and in contact-center shops average handle time and schedule adherence. 24/7 shops run shifts with a night and weekend differential. Deskside and field roles are onsite by definition, and those postings often carry a driver's license requirement and a stated lifting limit in pounds. |
| Typical hiring speed | Days to about two weeks from application to offer for staffing-agency and MSP roles. Public sector, healthcare and defense run far longer, and the delay is the background check, fingerprinting, drug screen or clearance rather than the interview panel. |
| Normal promotion window | 12 to 24 months on a busy tier-1 queue is enough to move to tier 2, desktop support or junior sysadmin, or to specialize into identity and Microsoft 365 administration, endpoint management, networking or a SOC. Help desk is the door most people without a degree walk through to reach IT work. It is a doorway, not a destination. |
What the job actually is, and the titles that mean the same thing
A help desk technician is the first human a user reaches when something at work stops working, and the person accountable for the ticket until it is resolved or handed to someone by name. The work splits into roughly three streams. Access and identity: password resets, account lockouts, MFA re-enrollment, group and license requests, new-hire setup, departure lockout. Endpoint and software: a laptop that will not join Wi-Fi or VPN, Outlook that will not sync, a stuck print queue, a failed update, a corrupted profile, imaging and re-imaging, application installs. Physical work: hardware swaps, docks and monitors, desk moves, conference-room AV, peripheral failures, depot RMA and shipping, asset tagging and e-waste.
Most of the contact arrives by phone, chat, email and web form, and in a lot of shops it is phone first. That is worth knowing before you accept, because a phone-first queue is measured. Ask in the interview which numbers the team is held to: first-contact resolution, time to resolve, reopen rate, and in contact-center environments average handle time and schedule adherence. Asking that question marks you as someone who has thought about the actual work rather than the job title.
The job is roughly one-third diagnostics and two-thirds communication. The user's description of the problem is frequently wrong. "The internet is down" usually means one application is unreachable. "It deleted my file" usually means the file was saved somewhere else. The skill being bought is converting an imprecise and often emotional complaint into a scoped technical question, and then telling the person what happened in words they can repeat to their own manager.
The second underestimated part is the written record. A ticket with no notes is a ticket nobody else can pick up, and your notes are the part of your work your manager and the tier-2 team actually see. Your fixes are invisible; your documentation is the evidence in the room when a promotion is discussed.
Titles vary more than the work does. Before you apply, decide which of these the posting is describing, because the resume you send should use the words that are in it.
- Help Desk Technician, IT Support Specialist, Service Desk Analyst, Technical Support Analyst: the same tier-1 job in most shops. Phone, chat, email and ticket queue, aiming at first-contact resolution.
- Desktop Support Technician, Deskside Support, IT Field Technician: hands on hardware, onsite, walking the floor. Usually tier 2, usually requires a driver's license if it covers multiple sites, and usually states a lifting requirement.
- Technical Support Engineer: supports a software product for external customers rather than employees for internal IT. Adjacent career, different skills, including reproducing bugs, reading logs and writing to engineering. Do not send one resume to both.
- NOC Technician: watches monitoring dashboards for circuits, servers and alerts rather than answering users. Shift work, often overnight, and a reasonable lateral entry if you prefer systems to people.
- Customer Support or Product Support at a BPO, meaning one of the large outsourced contact centers: fast to get hired, often remote, usually scripted, usually judged on handle time. It is real support experience and real income, and it builds less of the Windows, identity and endpoint evidence that a corporate IT hiring manager is looking for.
- MSP Technician or Tier 1 Engineer at a Managed Service Provider: the same tier-1 work across dozens of small client companies at once. High ticket volume, broad exposure, lower pay, and the fastest learning available to a beginner.
Four kinds of employer, four different hiring processes
Where you apply changes the process more than the job description does. Treat these as four separate campaigns rather than one.
If you have no experience, the realistic first offers come from MSPs, staffing-agency contracts, and public-sector or education help desks that hire on the credential plus a clean background check. Large-brand internal IT roles with a visible careers page draw heavy application volume for exactly the same work. Before you spend an hour tailoring for one of those, look at the applicant count the job board is already showing you, and make a rule: apply in the first days a posting is open, or do not bother.
There is a fifth route that is never advertised as an entry-level IT job, and almost nobody uses it deliberately: transferring from inside. Warehouses, hospitals, retail chains, school districts, universities, credit unions and manufacturers all have IT departments that would rather move a known quantity sideways than screen two hundred strangers. If you already work somewhere with an IT team, get your A+, tell your manager and the IT manager you are studying, volunteer for the next laptop refresh or rollout, and ask to be told when a tier-1 seat opens. That is the shortest path into IT available to someone with no IT experience.
- Internal corporate IT. You support your own colleagues. A recruiter screens, an IT manager interviews, a senior tech asks the technical questions. Deepest exposure to one environment (one Entra tenant, one Intune setup, one fleet standard), clearest promotion ladder, slowest hiring, heaviest competition on posted entry roles. A referral from anyone inside is worth more than another certificate.
- MSP. You support dozens of small client companies, sometimes more than a hundred. Hiring is fast, often a single conversation with the service manager plus a practical scenario, and they are explicitly buying throughput and attitude. The variety is structural rather than a promise: every client has a different fleet, firewall, mail setup and line-of-business application, so you touch problems an internal tech may wait years to meet. Pay is lower, utilization targets are real, and burnout is the common failure. Take it anyway if it is your first offer, because MSP tickets are the most transferable thing on a junior IT resume.
- Public sector, healthcare and education. City and county IT, school districts, universities, hospitals and clinics. Expect a formal application portal rather than email, a scored rubric, a panel interview with HR present, and a long gap before the start date for fingerprinting, a background check and sometimes a drug screen. Healthcare adds HIPAA training and immunization records; K-12 adds child-safety clearance. Pay bands are published, pensions still exist, and the job is more stable than anything else on this list. Apply through the portal exactly as instructed, because a strong resume sent outside the portal is not considered at all.
- Staffing agencies and contract-to-hire. TEKsystems, Insight Global, Robert Half, Apex Systems, Experis and their peers place a large share of entry-level IT support work, frequently as a six to twelve month contract with a conversion option, including the big fleet-refresh and migration projects. The recruiter is your screen, not the client. Build a relationship with two or three recruiters in your metro, answer their calls the same day, and be explicit about your commute radius and shift availability, because the recruiter is matching on logistics as much as on skills.
What actually gates the job: a credential, a check, and sometimes a vehicle
There is no license for this role. Nobody can stop you from doing the work, which means the gate is whatever the employer uses as a filter, and for entry-level help desk that is almost always a certification plus a background check.
CompTIA A+ is the one credential that functions as a key. It is not respected because it is difficult. It is respected because recruiters, HR filters and agency screeners all agree on what it means: that you know hardware, Windows, basic networking, mobile devices and the vocabulary of operational security. It is two exams, Core 1 and Core 2, and you are not A+ certified until both are passed. CompTIA retires and renumbers the pair roughly every three years, so verify the live exam codes and download the current objectives PDF from CompTIA before buying a course. A study bundle written for the retired pair is the most common wasted purchase in this field.
Before you pay list price for anything, find out whether someone else will pay. Your state's WIOA-funded workforce program through an American Job Center will cover the whole cost of training and the exam voucher in many states. A community college course often bundles the voucher. CompTIA sells discounted academic vouchers. IT support apprenticeships are real and are listed on apprenticeship.gov, and they pay while you train. Nonprofit programs such as Per Scholas, NPower and Year Up run free training that ends in placement with employer partners. If you are leaving the military, look at SkillBridge and at your branch's credentialing assistance. None of this is secret; it is just unevenly used.
Everything else is optional and should be chosen for the kind of shop you are targeting rather than collected for its own sake. Two certificates plus one real environment you administer beats five certificates and nothing to show.
The non-technical gates are the ones that surprise people, and they are what delays a start date. Clear them early.
- CompTIA A+: the default entry key. Get this first unless you already have IT work experience to point at.
- ITIL 4 Foundation (PeopleCert): teaches the vocabulary of service management, including the difference between an incident, a service request, a problem and a change, what an SLA actually commits to, and why a workaround is logged separately from a fix. One exam, no lab. Disproportionately useful at large enterprises and MSPs, where using these words correctly marks you as someone who has worked a real queue.
- Microsoft MS-900 (Microsoft 365 Fundamentals) and MD-102 (Endpoint Administrator): MS-900 is a weekend. MD-102 is the one that maps directly onto the daily work of a Windows shop in 2026, covering Intune, Autopilot, compliance and configuration policies, and Entra-joined devices. AZ-900 is cheap credibility if the posting mentions Azure. SC-300 is the next step if you find you like identity work.
- CompTIA Network+ and Security+: Network+ earns its keep if the posting mentions VLANs, 802.1X, switches or site networking. Security+ is less about help desk and more about your next job, since it is the standard filter for SOC analyst roles and for defense-contractor positions under the DoD 8140 framework. Do not take either before A+ without a specific reason.
- Google IT Support Professional Certificate (Coursera) and HDI's support-center certifications: both legitimate, both weaker signals than A+ on their own. The Google certificate is a good structured on-ramp for someone with no technical background and is best treated as study material that ends in an A+ attempt.
- Mac and Linux depth when the fleet needs it: plenty of design, media, education and tech-company fleets are half Mac or more. Jamf's 100-level training is free, and knowing Apple Business Manager, automated device enrollment and the real difference between a Mac and a Windows configuration profile puts you ahead of most A+ holders, who learned very little about macOS.
- Background check, fingerprinting, drug screen, driving record. Healthcare, finance, government, K-12 and defense all run checks, and they are the long pole in the process. If you have a record, find out now whether it is disqualifying for that sector rather than after an offer.
- Clearance, if you are aiming at defense work. You cannot apply for a clearance yourself; an employer sponsors it, and the posting will say whether you must be clearable for public trust or Secret. Being clearable now is worth money later, and the processing time is measured in months.
- A vehicle, sometimes. Deskside and multi-site field roles routinely require a valid driver's license, sometimes your own insured car with mileage reimbursed. Postings often state a lifting requirement in pounds, because the job includes carrying equipment.
How hiring works in 2026-27, stage by stage
Two things about this market are worth stating separately, because merging them produces a tidy story that is wrong. First, AI-assisted applying means a single posting draws far more applicants than it did three years ago, and many of them look tailored, because they were tailored by a model. Second, the easiest slice of the ticket mix really is being absorbed at companies that have deployed self-service and virtual agents, which has thinned the number of pure queue-answering seats while leaving demand for techs who can do the whole job intact. Both facts raise the bar on evidence. Neither has closed the door.
What cuts through the volume is specificity that a general-purpose rewrite cannot fake. "Provided excellent technical support in a fast-paced environment" reads as generated whether or not it was. A line in this shape reads as lived: logged 340 tickets in six months in Freshservice for a 180-user nonprofit, re-imaged 22 laptops with Intune and Autopilot, wrote the knowledge article that stopped the VPN password-expiry calls. Use your own real numbers, whatever they are. The first screen is often a model summarizing your resume for a recruiter who then reads the summary, which rewards plain structure, explicit product names and numbers with units, and punishes design-led two-column templates that scramble reading order.
One concrete thing to aim at. Windows 10 reached end of support on 14 October 2025, and paid Extended Security Updates are time-limited and priced to push migration, so Windows 11 refresh and migration projects are still generating contract deskside work through 2026 and into 2027. These projects hire in batches through staffing agencies, they are openly short-term, and they are the easiest first line of verifiable enterprise experience a beginner can get. Ask recruiters directly what refresh or migration projects they are staffing, and make sure the words imaging, Autopilot and migration are on your resume if you have done any of it, even in a lab.
Speed is the other thing to understand. For agency and MSP roles the whole process can take four days. Answer your phone, reply to email within hours, and brief your references in advance. A responsive candidate on Tuesday beats a better candidate who replies on Friday, because the requisition is already filled.
The stages, where they exist:
- Recruiter or agency screen, 15 to 30 minutes, by phone. This is vocabulary and logistics: which ticketing system, which OS mix, shift availability, commute radius, pay expectation, certification status, work authorization and willingness to be checked. Have a 60-second version of who you support, what you fix and what you are studying. Name a number for pay, taken from BLS OES and local pay-transparency postings, stated calmly. Ask whether the role is W-2 contract, contract-to-hire or direct.
- Hiring-manager interview, 30 to 60 minutes, with an IT manager or service desk lead. Graded on reliability and manner as hard as on knowledge: will you show up for a 7am shift, will you stay calm with an angry executive, will you document. Expect direct questions about attendance and schedule, and answer honestly, because accepting a shift you cannot sustain is how new techs lose the job in month two.
- Technical conversation or practical, 30 to 60 minutes, with a senior tech or tier-2 lead. Usually spoken scenarios ("a user says their computer is slow"), sometimes a short written quiz, occasionally a real machine or a shared screen with a broken VM. Graded on your process and your questions rather than on recall of port numbers.
- Peer or roleplay stage, where it exists. Someone plays a frustrated non-technical user. Graded on whether you drop the jargon, confirm the symptom, set an expectation and close the loop. Some contact centers replace this with a typing and CRM-navigation test.
- References and the check. Two work references who will actually answer the phone, ideally including someone who can speak to your attendance. Then the background check, fingerprinting or drug screen: days for private sector, weeks to months for government, healthcare and cleared work.
- Offer. Hourly rate, shift and differential, on-call expectations, whether the role is W-2 contract or direct, and the conversion terms if it is contract-to-hire. Ask what the tier-2 path looks like and whether certification exams are reimbursed. Both answers tell you how long you should expect to stay.
Proof when you have never held an IT job: the lab and the ticket log
The hiring manager's real fear is not that you lack knowledge. It is that you have never had a stranger shout at you about a laptop and have never touched a production system you could break. Everything below exists to answer that fear with evidence instead of enthusiasm.
Build the lab, then do work in it. A screenshot of an installed operating system proves nothing. A short write-up saying "created an OU per department, applied a GPO that mapped a department drive, then deliberately broke the mapping and traced it with gpresult" proves you can diagnose. The hardware cost is close to zero: an old desktop or a used laptop with 16GB of RAM, a free hypervisor, and the evaluation and trial tiers the vendors publish.
Then fix computers for people who are not you. This is the part almost nobody does, and it is the part that converts interviews. A local nonprofit, a church, a community center, a small clinic, a youth sports league, a library, a volunteer computer-refurbishing project: all of them have broken printers, shared-drive permission messes, nobody running updates, and no budget. Do it properly. Stand up a free ticket system, make them email it instead of texting you, log every request, write a knowledge article for anything you fix twice, and keep a running count. Six months of that is genuine support experience with a real reference attached, and you can describe it on a resume without one word of exaggeration.
Present it correctly. Assume the hiring manager will not click a portfolio link, so the resume itself has to carry the numbers and the link is backup. Keep a one-page public write-up, a plain GitHub README or a simple site, with eight to twelve entries in a fixed shape: the symptom as the user reported it, what you checked and ruled out, the actual cause, the fix, and how you would prevent it. Bring it to the interview on paper or on a tablet. When they ask you to describe a problem you solved, you open it instead of groping for a story.
- A Windows domain from scratch. A Windows Server evaluation ISO in Hyper-V, Proxmox or VirtualBox as a domain controller with DNS and DHCP, and a Windows 11 client joined to it. Then do the tier-1 work: create and disable users, reset passwords, build organizational units, apply a Group Policy Object, set up a shared folder, and prove you understand NTFS versus share permissions by making them disagree on purpose.
- A cloud tenant you administer. A Microsoft 365 trial tenant with Entra ID, licenses assigned to test users, MFA enforced, one conditional-access policy, and a security group that drives a license. The Microsoft 365 developer sandbox is now harder to qualify for, so check the current eligibility before you count on it. Write "Entra ID" on your resume, never "Azure AD", because Microsoft renamed it in 2023 and the old name quietly dates you.
- Endpoint management on your own machine. Enroll a device in an Intune trial, push a compliance policy, deploy one application, run a remediation script, then wipe and re-enroll it. If you are aiming at a Mac-heavy fleet, do the Jamf or Kandji equivalent and learn what automated device enrollment actually does.
- A ticketing system you have actually used. Jira Service Management's free tier covers a small number of agents; Freshservice, Zoho Desk and the open-source options (osTicket, GLPI, Zammad) all work. Configure a queue, set priorities and an SLA, and log your own work through it. Being able to say "I have worked a queue in Jira Service Management" removes an objection that stops a lot of beginners.
- Asset and inventory discipline. Snipe-IT is free. Track fifteen real devices with serial numbers, assignment, warranty dates and disposal. Inventory is the unglamorous task every understaffed IT department is behind on, and offering to own it is how a new tech becomes useful fast.
- Networking you can show on a diagram. A pfSense or OPNsense box, two VLANs, a static reservation, a DNS override, and a written trace of what happens when a client cannot reach a host. Then be able to run and explain ipconfig, nslookup, ping, tracert and a DNS flush without reading from a list.
- A written knowledge base. Five to ten articles in the format a real service desk uses: the title phrased the way a user would search it, symptoms, prerequisites, numbered steps, screenshots, and what to do if the steps fail. This single artifact has become a hiring signal in its own right, for the reason in the AI section below.
The resume: what a 20-second screen reads, and what it throws away
Your resume will be skimmed by a recruiter or an agency screener matching the posting's nouns, often with a model summarizing it first. Write a plain single-column document in a standard font, use ordinary section headings (Summary, Certifications, Technical Skills, Experience, Education), save as PDF unless the portal demands .docx, and put the ticketing system, the OS mix and the identity platform where they cannot be missed. No graphics, no skill-rating bars, no columns that interleave when parsed, no headshot.
Lead with the credential and the proof rather than an objective. Two lines naming your A+ status, the environment you have worked in even if it was voluntary, and the user count is enough. Then a Certifications block, because that is the line being searched for.
Quantify with units a manager can calibrate: tickets per week, users supported, sites covered, devices imaged, first-contact resolution, average time to resolve, CSAT, backlog reduced, articles written. If you have no paid IT job, those numbers come from the volunteer work and the lab. Label them honestly as such and they still count. Inflation is the failure mode: a panel discounts a wild claim and does not discount an honest small one.
- Name the products, with their current names. Entra ID (not Azure AD), Intune, Autopilot, Microsoft 365 and Exchange Online, Active Directory and Group Policy, Jamf or Kandji, ServiceNow or Jira Service Management or Freshservice, Okta, Google Workspace admin, TeamViewer or ScreenConnect or Quick Assist, your VPN client by name, Windows 11 and macOS versions, PowerShell.
- One line per ticket type you can own, with a number in front of it. "Resolved around 60 tickets a week at first contact across 400 users (Windows 11, Microsoft 365, Entra ID, Intune)" tells a manager more than a paragraph of adjectives.
- Keep prior non-IT work and translate it. Retail, hospitality, call center, military, trades and teaching all prove the thing managers actually worry about: showing up, staying calm, working a queue, following a procedure, de-escalating. "Handled 200+ customer interactions per shift; covered the opening shift for two years with no missed days" is a relevant line on a help desk resume. Delete the duties that are not relevant and keep the volume, the reliability and the conflict handling.
- State shift and travel availability explicitly if you are flexible. "Available for weekend and overnight shifts; valid driver's license, own vehicle" removes the exact doubt that kills tier-1 candidates.
- Federal and public-sector applications are the one exception to the one-page rule. A USAJOBS resume is long on purpose: for each job give month and year, hours per week, and enough duty detail for an HR specialist to score you against the announcement's required experience, then answer the questionnaire literally rather than modestly. A tight private-sector one-pager gets scored as experience not demonstrated.
- Thrown away: the objective statement, "excellent communication skills", "hardworking team player", a 30-item skills grid of things you have read about, your high-school GPA, an unrelated degree placed above the certifications, and a two-page document for a first private-sector job.
- Thrown away: duties with no object. "Provided technical support to end users" says nothing. Which systems, how many users, which problems, resolved or escalated?
- Dangerous: claiming tools you cannot discuss. If ServiceNow is on your resume you will be asked what a change request looks like in it. Listing a tool you have only watched a video about is the fastest way to lose a technical screen.
- Two documents rather than one if you are applying to both internal IT and product support. Internal IT wants Windows, identity, endpoint and hardware. A software vendor's support team wants log reading, reproduction steps, APIs and written clarity. Change which bullets lead.
The interview: the scenario, the roleplay, and the identity-verification question
Help desk technical interviews are mostly spoken scenarios, and they are graded on process. The interviewer is listening for whether you ask scoping questions before proposing a cause, whether you change one thing at a time, whether you know when to stop and escalate, and whether you say what you would write in the ticket. A confident wrong guess scores worse than a methodical "here is what I would check first, and why".
Learn one visible framework and use it out loud: confirm the symptom in the user's words, establish scope (one user or many, one device or all, wired or wireless, when did it last work, what changed), form one theory, test it, document, then either resolve or escalate with everything the next person needs. Saying the steps aloud is the point, because nobody can grade thinking they cannot hear.
Two questions deserve specific preparation, because candidates reliably fumble them. The first is the identity-verification question below, which has become a standard screen at security-conscious employers. The second is "tell me about a time you did not know the answer", where the correct answer involves a bounded attempt, a time limit, telling the user what is happening, and an escalation with notes. Never "I keep going until I solve it". Unbounded heroics on a tier-1 queue are a liability and interviewers know it.
Expect, in some form:
- "A user says the internet is down." Graded on scoping, not on an answer. Is it only them? Wired or Wi-Fi? Can they reach anything at all, or only one application? Did it work this morning, and what changed? Can other people in the same room reach it? Only then: ipconfig, a DNS check, VPN state, the Wi-Fi profile, and whether the site or service itself is down. Your first move is a question and your second is working out whether this is an incident affecting many people, because that changes who needs to know in the next five minutes.
- "A user cannot print." The classic, because it has a dozen causes in layers: the wrong printer selected, a stuck queue, the driver, credentials for a secure-release printer, the network path, offline status, the physical device. Walk the layers out loud and say where you would stop and go look at it in person.
- "A caller says they are an executive, they are locked out, they are about to go on stage, and they are calling from a number that is not in the directory." This is the security question and it is now asked deliberately. The right answer: urgency is itself a flag, you follow the documented verification procedure every time and for everybody, you do not reset credentials or re-enroll MFA on the strength of a voice, and you escalate rather than make an exception, ideally verifying out of band through a channel already on record, a manager, or an in-person or video check against the record. Help desk reset paths are a known and heavily used intrusion route: the CISA and FBI joint advisory on the Scattered Spider group, AA23-320A, describes this exact technique, and it is how the widely reported 2023 casino-operator intrusions began. "I would make an exception for a VIP" fails the interview outright at any shop that has read that advisory.
- "A user's machine is slow." Graded on whether you resist the reflex to reimage. Since when, what is running, disk and memory pressure, pending updates, encryption or indexing still in progress, a failing drive, the age of the hardware, and whether it is one machine or the whole floor. Then: what would you tell the user while you check, and what do you write in the ticket.
- "Walk me through setting up a new hire." Account creation and group membership, license assignment, MFA enrollment, device enrollment and imaging, application access, hardware and peripherals, first-day orientation, and what you check before you call it done. A good answer mentions that you follow the onboarding checklist and update it when it turns out to be wrong.
- A basics round. DNS versus DHCP in one sentence each, what an IP conflict looks like, the difference between NTFS and share permissions, what Group Policy does, what safe mode is for, and how you would prove whether a problem belongs to the device or the account (log the user in on another machine; log another user in on that machine). Short, correct, no padding.
- The reliability questions, asked plainly. What shift can you work, how far is your commute, how do you get here in bad weather, have you ever missed a shift. Tier-1 teams are hurt more by absence than by ignorance, so these are not small talk. Answer concretely and do not overpromise.
Pay, shifts, and what to do in the first 18 months
For pay, go to the source rather than to a salary-aggregator average. The occupation is Computer User Support Specialists, SOC code 15-1232, in the BLS Occupational Employment and Wage Statistics, which publishes median and percentile wages nationally, by state and by metropolitan area. That is the only free figure that is both local and defensible in a negotiation. Treat 15-1231, Computer Network Support Specialists, as the adjacent and generally better-paid code you are aiming at next. Then calibrate against live postings in pay-transparency states, where a range is legally required in the ad; filtering your search to Colorado, Washington, California, New York, Illinois or Minnesota gives you current real numbers even if you work somewhere else.
Three things move the rate more than another certificate does. Sector: finance, defense, biotech and specialized healthcare IT pay above retail, hospitality, nonprofit and small-business MSP work for the same tasks. Shift: nights, weekends and on-call carry a differential, and a willingness to take the unpopular shift is the cheapest raise available to a beginner. Scarcity of a specific stack: Mac fleet depth, clinical application support, a clearance, Spanish or another language your user base speaks, or an industrial or lab environment all command a premium, because the supply of help desk techs who have them is small.
Contract-to-hire deserves clear eyes. A W-2 contract through a staffing firm often pays a competitive hourly rate with thin or no benefits, and conversion is a real but not guaranteed outcome. It is usually still the right first job, because it starts the clock on verifiable experience, and six months of logged tickets in an enterprise environment changes what you can apply for. Ask at offer time what proportion of contractors on that team converted, and how long it took.
Take the first offer that puts you on a real queue with real systems, then be deliberate for 18 months. Months one to three: learn the environment and the escalation path, read the knowledge base, and start writing articles for the gaps you hit. Months three to nine: pick up the tickets nobody likes, meaning printers, imaging, inventory, onboarding checklists and the Mac users, because unclaimed work is the fastest route to being the person who owns something. Months nine to eighteen: automate part of your own queue, write down what it saved, and get the second certification pointed at where you want to go.
Pick that direction on purpose, because the branches diverge sharply. Identity and Microsoft 365 administration (MD-102, then SC-300) is the shortest path out of the queue in most Windows shops. Endpoint management is the same bet with a Jamf or Intune specialism attached. Networking (Network+, then CCNA) leads to network technician and network engineer. Security (Security+, then a SOC analyst role) is the most competitive and most hyped branch: real, but it does not skip the queue, and SOC teams hire help desk veterans precisely because they have handled incidents under pressure. Cloud (AZ-104 or the AWS equivalent) pays well and wants systems fluency first. Ask your employer whether exams are reimbursed before you pay for any of them. Many reimburse, and almost nobody asks.
One honest caution. Help desk is an excellent doorway and a poor plateau. The ceiling on tier-1 pay is low, the work does not get easier with seniority, and techs who stay five years without specializing usually find the next move harder than it was at eighteen months. Go in knowing which room you intend to walk into.
What a help desk technician has to know about AI in 2026-27
The honest version: AI has not replaced the help desk and it has not changed the core of the job, which is still a person who is stuck and a device that is not behaving. What changed is the shape of the queue and the content of some tickets. Self-service password reset, virtual-agent deflection and suggested-resolution panels have absorbed a slice of the simplest repeat contacts at the companies that deployed them, so the tickets reaching a human skew harder and more ambiguous. At the same time a new ticket category arrived, the users' own AI tools, and the help desk inherited it by default.
Three practical consequences, and they are the ones to speak to in an interview. First, deflection quality is knowledge-base quality, because a virtual agent answers from the articles your team wrote; writing good articles became measured output instead of a chore. Second, the help desk is now the control point for shadow AI, because employees paste work into whatever assistant they can reach and you are the person they ask about it. Third, identity verification got harder in a specific, documented way: a convincing voice or a plausible video call is cheap to produce, and the help desk reset path is a known intrusion route, spelled out in the CISA and FBI advisory on the Scattered Spider group, AA23-320A.
Be careful about the opposite error too. A great deal of what is sold as AI in service management is ticket auto-categorization, summarized transcripts and a suggestion panel beside the queue. Useful, unglamorous, and frequently wrong in ways you are expected to catch. If an interviewer asks what you think of AI on the service desk, the credible answer is specific: name the tool, say what it genuinely absorbs, say where it fails, and say who owns the ticket when it fails. You do.
What follows is what to be able to show. None of it requires writing software.
Writing knowledge articles that a virtual agent can answer from
Every deflection layer, whether it is ServiceNow's assistant, Zendesk's AI agents, Freshservice's, or a Moveworks-style virtual agent, answers out of the knowledge base. Thin, stale or hedged articles produce confidently wrong bot answers and a worse ticket afterwards. Service desk managers now treat article authorship as core output, and a tier-1 tech who writes well is visible to management in a way a fast closer is not.
Show it: Bring five articles you wrote, in service-desk shape: title phrased the way a user would actually search it, symptoms, prerequisites, numbered steps, screenshots, and an explicit failure branch. Say which one you wrote because you had answered the same question four times, and what happened to those calls afterwards. If you have no job yet, write them for the volunteer environment you support, because the artifact is identical.
Supporting the AI tools your users now have
Microsoft 365 Copilot, Gemini in Google Workspace, ChatGPT Enterprise and their peers generate real tickets: the assistant cannot see a file, it returns something the user believes is wrong, the license is not assigned, the feature is not enabled for their group, or it surfaced a document the user is surprised they can open. That last one is a permissions and oversharing problem wearing an AI costume, and it lands on the help desk first.
Show it: Be able to explain in plain language why an assistant "cannot see" a document (it reaches only what the signed-in user already has access to, and only once the content is indexed), where license assignment lives, and what you would hand to the identity or security owner rather than fix yourself. Say out loud that you would never widen a permission to make an AI answer work.
Identity verification that survives a cloned voice
The highest-consequence thing a tier-1 technician does is prove who is on the other end before changing credentials or re-enrolling MFA. Voice and video impersonation is cheap now, and attackers specifically target help desks with urgency, seniority and a plausible story. This is the question that separates candidates at security-conscious employers.
Show it: Describe a procedure rather than an instinct: the documented identity check applied to everyone including executives, out-of-band confirmation through a channel already on record, manager or in-person confirmation for high-privilege accounts, a hard rule that urgency never shortens the check, and logging the attempt when something feels wrong. Name the attack by name. Then say the sentence interviewers want to hear: you would rather be the tech who annoyed a vice-president than the tech who onboarded an intruder.
Working alongside AI triage without outsourcing your judgement
Modern ITSM platforms auto-categorize tickets, summarize phone and chat transcripts, and suggest a resolution from similar past tickets. The suggestions are sometimes right, often adjacent, and occasionally dangerous, such as a fix drawn from a different OS version or a different client tenant. Your name is on the resolution either way.
Show it: Give a concrete instance where you checked a suggestion before acting: you compared the proposed fix against the vendor documentation or the device's actual configuration and found it did not apply. State your rule: you do not run a command you cannot explain, and you do not close a ticket on a summary you have not checked against what the user actually said.
Using an assistant for diagnosis without leaking the company's data
Pasting an error, a log excerpt or a configuration into a public chatbot is the most common data-handling mistake junior technicians make, and it is exactly the behavior security teams are watching for. The skill is doing it safely rather than abstaining, because an assistant is genuinely good at decoding an obscure error code or drafting a clear user-facing explanation.
Show it: Say which tool you would use and why it is the approved one, that you strip user names, hostnames, IP addresses, serial numbers, ticket contents and anything identifying a named person before asking, and that you verify the answer against the vendor's documentation before running it on someone's machine. Mention the acceptable-use policy as the thing you would read in week one.
Light scripting, now reachable from tier 1
An assistant has put basic PowerShell, Microsoft Graph queries and Intune remediation scripts within reach of a technician who is not a developer. This is now one of the clearest differentiators between a tech who stays on the queue and one who gets the tier-2 job, because automating repeated work is visible and measurable.
Show it: Three small scripts you actually run, with the time they save and, more important, what you did to make them safe: tested against a pilot group first, ran with -WhatIf before committing, had a rollback, and had someone senior read it before it touched real accounts. Being able to read a script line by line and explain it matters more than having written it unaided, so say plainly that an assistant helped, then demonstrate that you understand every line.
Measuring deflection and automation, and saying what it cost
Service desk managers are now asked to show that their self-service and virtual-agent spending reduced contact volume. A tier-1 tech who can talk about deflection rate, reopen rate, first-contact resolution and the places where automation created more work than it saved is speaking the manager's language, which is how you get invited into planning conversations.
Show it: Pick one repeat ticket type you removed, such as a password-expiry wave, a VPN prompt or a printer-driver reinstall, and give the before and after: contacts per week beforehand, what you changed (an article, a self-service flow, a group policy, a pilot), contacts per week after, and one case where the automated path failed and how you caught it.
Being clear about what AI has not touched
Overclaiming disruption reads as naive to someone who runs a service desk. Hands on hardware, imaging and re-imaging, desk moves and conference-room AV, day-one setup for a new hire, the user whose description of the problem is wrong, the outage judgement call about who to wake up, and the executive who wants a person standing next to them: none of that has been automated, and none of it is close.
Show it: When the AI question comes, answer in two halves: here is what it genuinely absorbed on the queues I have seen, and here is the part of this job that is still a person with a screwdriver and a calm voice. That balance is the answer that gets hired. A candidate who claims AI will run the service desk and a candidate who claims it changed nothing both sound like they have never worked one.
What a screen is looking for
These are the terms that a resume screen, human or automated, is matching against for this role. Use the ones that are true of you, in the words the posting uses.
- Help desk technician
- IT support specialist
- Service desk analyst
- Desktop support technician
- Tier 1 support
- Tier 2 escalation
- First-contact resolution (FCR)
- Ticket queue management
- SLA adherence
- Average handle time (AHT)
- Schedule adherence
- Incident management
- Service request fulfillment
- Problem management
- Change request
- Knowledge base article authoring
- Knowledge-Centered Service (KCS)
- CSAT
- CompTIA A+
- CompTIA Network+
- CompTIA Security+
- ITIL 4 Foundation
- Microsoft MS-900
- Microsoft MD-102 Endpoint Administrator
- Microsoft AZ-900
- Microsoft SC-300
- Google IT Support Professional Certificate
- HDI Support Center Analyst
- Windows 11
- Windows 11 migration
- Windows 10 end of support
- macOS support
- Active Directory
- Group Policy (GPO)
- Organizational units
- Microsoft Entra ID
- Conditional Access
- Multi-factor authentication (MFA) enrollment
- Self-service password reset (SSPR)
- Microsoft 365 administration
- Exchange Online
- Outlook troubleshooting
- Microsoft Teams support
- Google Workspace admin
- Okta
- Microsoft Intune
- Windows Autopilot
- Configuration Manager (SCCM)
- Jamf Pro
- Kandji
- Apple Business Manager
- Device enrollment
- Laptop imaging and re-imaging
- Hardware troubleshooting and repair
- Printer and print queue support
- VPN troubleshooting
- DNS
- DHCP
- TCP/IP
- VLAN
- 802.1X
- Wi-Fi troubleshooting
- ServiceNow
- Jira Service Management
- Freshservice
- Zendesk
- Ivanti
- ManageEngine ServiceDesk Plus
- Remote desktop support
- Intune Remote Help
- TeamViewer
- ScreenConnect
- Quick Assist
- PowerShell
- Microsoft Graph
- Asset management
- Snipe-IT
- Onboarding and offboarding
- Identity verification procedure
- Social engineering awareness
- Phishing response
- Microsoft 365 Copilot support
- AI virtual agent deflection
- Shift work and on-call rotation
- Contract-to-hire
- Managed service provider (MSP)
- HIPAA awareness
- GS-2210 IT specialist
- Background check and fingerprint clearance
Mistakes that cost people this job
Collecting certifications instead of touching systems. Three CompTIA certs, no lab, no users ever supported, and no answer when a senior tech asks what you actually did with Group Policy.
One certification plus one real environment beats three certifications and nothing. Get A+, then build the domain, the tenant and the ticket queue, then support real people somewhere unpaid. The certificate gets your resume read; the lab and the ticket log win the interview.
Applying only to fully remote roles.
Remote tier-1 postings compete nationally against candidates who already have enterprise experience, and the hands-on half of this job cannot be done from a bedroom anyway. Apply onsite and hybrid first and say so on the resume: deskside and multi-site work is where beginners get hired. Remote becomes realistic after a year of logged tickets.
Turning down the MSP job or the agency contract because it is not a "real" role or pays below the aggregator average.
Take the first offer that puts you on a live queue with production systems. Every MSP client has a different fleet, firewall and line-of-business application, so the exposure is wide by design, and a six-month contract converts your resume from aspirational to verifiable. Negotiate the shift and the conversion terms, not the existence of the job.
Answering the VIP lockout question by making an exception: "if an executive is in a hurry I would reset it and sort out the paperwork after."
Say that urgency is a flag, that the documented verification procedure applies to everyone, and that you would confirm out of band through a channel already on record before changing any credential or MFA enrollment. Help desk resets are a documented intrusion route. Being briefly unpopular is the correct outcome.
Jumping to a cause before scoping the problem. "The internet is down." "I would reboot the router."
Ask first, out loud: only you or others, wired or wireless, everything or one application, when did it last work, what changed. Then one theory, one test, one change at a time, and a note in the ticket. Interviewers are grading the questions, not the guess.
A designed resume, with two columns, skill bars, icons and a headshot, because it looks professional.
Single column, standard headings, plain font, PDF. The first reader is often a parser or a model producing a summary, and a layout that scrambles reading order can lose your certifications entirely. Spend the effort on numbers with units instead of on typography.
Listing tools you have only watched a video about, then being asked to describe using one. ServiceNow on the resume and no idea what a change request looks like in it.
List only what you can discuss for two minutes. Split the skills block honestly if you need to: "Used daily" versus "Home lab and studying". Nobody is penalized for the second heading; everybody is penalized for being caught out under the first.
Writing "Azure AD" or "Office 365", or naming an operating system two versions behind. Small wording, immediate dating.
Entra ID, Microsoft 365, Windows 11. If you worked a Windows 10 fleet, say so as the migration work it was: Windows 10 reached end of support on 14 October 2025, and "migrated 180 endpoints to Windows 11" is a strong line in 2026 and 2027.
Describing fixes as actions with no diagnosis. "Restarted the machine and it worked." "Reimaged it."
Name what you ruled out and why the fix followed. "The slowness was pending disk encryption, confirmed in the encryption status, not the hardware, so I explained the timeline to the user instead of reimaging a machine that was fine." Diagnosis is the skill being bought.
Treating the shift and commute questions as formalities, then accepting a 6am start or a 70-minute drive you cannot sustain.
Answer honestly at the screen. Absence hurts a tier-1 team more than inexperience does, managers have been burned by it, and a candidate who states their real constraints is more hireable than one who agrees to everything and quits in month two.
Treating AI as either the end of the job or as irrelevant. Both answers mark you as someone who has never worked a queue.
Be specific. Self-service and virtual agents have absorbed a slice of the simplest repeat contacts; hardware, onboarding, misdescribed problems and outage judgement have not moved. Then say what you can do about it: write the articles the bot answers from, and automate one repeat ticket type and measure it.
Ignoring the sector gates until after the interview: the fingerprint check, the drug screen, the clearance sponsorship, the driving record, the immunization record.
Find out what your target sector requires before you apply. Public-sector, healthcare, K-12 and defense offers stall on paperwork, not on panels. If a check is going to be a problem for you, learn that now and aim at sectors that do not run it.
Questions people ask
Can you get a help desk job with no degree and no experience?
Yes, and help desk remains the widest no-degree entrance to IT. The US Bureau of Labor Statistics does not list a bachelor's degree as the typical entry-level education for computer support specialists, and OPM's qualification standard for the federal GS-2210 IT series qualifies candidates on IT-related experience and demonstrated competencies rather than on a degree alone. What replaces the degree is a credential plus evidence: CompTIA A+, a home lab you have done real administrative work in, and ideally six months of unpaid support for a nonprofit, clinic, school or community organization with a ticket log and a reference attached. Managed service providers, staffing-agency contracts and public-sector help desks hire on that combination routinely. Large-brand internal IT postings are the hardest first door, not because of the degree but because of application volume.
Which certification should I get first for help desk?
CompTIA A+, unless you already have IT work experience to point at. It is two exams, Core 1 and Core 2, and you are certified only after passing both. Each attempt costs a few hundred US dollars at list price, and CompTIA renumbers the pair roughly every three years, so check the live exam codes and the current objectives PDF on CompTIA's site before buying any study material. Look for a funded voucher first through a state WIOA program, a community college course or CompTIA's academic pricing. After A+, choose the second certificate by the shop you are targeting rather than by prestige: ITIL 4 Foundation for large enterprises and MSPs, Microsoft MD-102 for a Windows and Intune fleet, Network+ if the posting mentions VLANs or switching, Security+ only if you are aiming at a SOC or at defense-contractor work.
Is CompTIA A+ still worth it in 2026?
Yes, as a key rather than as an education. Its value is that recruiters, HR filters and agency screeners all agree on what it means, so it gets your resume read. It will not by itself get you hired, because most other entry candidates have it or are studying for it, and it teaches very little about the cloud identity and endpoint management work that fills a modern help desk day. Pair it with a Microsoft 365 and Entra ID tenant you actually administer and an Intune trial you have enrolled a device into. The agreed credential plus current-stack evidence is what separates two otherwise identical beginners.
How long does it take to get a first help desk job?
Plan on three to six months from a standing start: roughly two to four months of part-time study for both A+ exams, about a month of evenings building and documenting the lab, and several weeks of applying. The application stage itself is fast once you are in it, and staffing-agency or MSP processes can run from first call to offer inside a week. The long delays are sector-specific: government, healthcare, K-12 and defense offers wait on fingerprinting, background checks, drug screens or clearance sponsorship, which can add weeks or months after the panel has already decided to hire you.
Is help desk being replaced by AI?
No, but the ticket mix has changed and the entry bar moved with it. Self-service password reset and virtual-agent deflection have absorbed a slice of the simplest repeat contacts at the companies that deployed them, so there are fewer purely queue-answering seats and the tickets reaching a human are harder and more ambiguous. Nothing has automated hands-on hardware, imaging, day-one setup for a new hire, desk and conference-room work, the user whose description of the problem is wrong, or the judgement call about who to wake up during an outage. Two things did get more valuable: writing the knowledge articles a virtual agent answers from, and verifying identity carefully, because voice and video impersonation of staff is cheap now and help desk reset paths are a documented intrusion route.
What does a help desk interview actually test?
Process, manner and reliability, in that order, and far less trivia than candidates expect. You will get spoken scenarios: the internet is down, the printer will not print, the machine is slow, walk me through setting up a new hire. The grader is listening for scoping questions before a theory, one change at a time, knowing when to escalate, and saying what you would write in the ticket. Expect a de-escalation element, sometimes as a roleplay with a frustrated user. Expect an identity-verification question about a locked-out executive calling from an unknown number. And expect blunt questions about shift availability, commute and attendance, because absence damages a tier-1 team more than inexperience does.
What should a help desk resume include if I have no IT job yet?
Single column, standard headings, PDF. A two-line summary naming your certification status and the environment you have worked in. A Certifications block high on the page. A Technical Skills block in current product names: Entra ID rather than Azure AD, Microsoft 365, Intune, Windows 11, macOS, and the ticketing system you have used. Then experience bullets with numbers and units drawn from volunteer support and the lab, labelled honestly as such: tickets logged, users supported, devices imaged, articles written, systems administered. Keep prior retail, hospitality, call-center, military or trades work and translate it into queue volume, shift reliability and conflict handling, which is exactly what the hiring manager is worried about. Delete the objective statement and every adjective about yourself. Federal applications are the exception: a USAJOBS resume should be long and detailed, with months, hours per week and duties scored against the announcement.
How much do help desk technicians get paid?
Use the US BLS Occupational Employment and Wage Statistics under SOC code 15-1232, Computer User Support Specialists, which publishes median and percentile wages nationally and by state and metro area. That is the figure you can defend in a negotiation. For current live numbers, filter job searches to pay-transparency states such as Colorado, Washington, California, New York, Illinois and Minnesota, where postings must carry a range. The work is usually hourly and non-exempt, so overtime is paid, and nights, weekends and on-call generally carry a differential. Sector moves the rate more than certifications do: finance, defense, biotech and specialized healthcare IT pay above nonprofit, retail and small-business MSP work for the same tasks.
Should I take an MSP or a contract job as my first IT role?
Usually yes. An MSP puts you across many client environments at once, each with its own fleet, firewall and line-of-business application, so the range of problems you see is wide by design, and a staffing-agency contract converts your resume from aspirational to verifiable within six months. The honest trade-offs: MSP pay is lower, utilization targets are real, and burnout is the common failure, while contract roles often carry thin benefits and conversion is a possibility rather than a promise. Ask at offer time what proportion of contractors on that team converted and how long it took, keep your own running count of tickets and systems from day one, and set yourself a review date rather than drifting.
How do I move from help desk into sysadmin, cloud or security?
Give it 12 to 24 months on a busy queue, then move deliberately. Claim the work nobody wants, meaning imaging, printers, inventory, onboarding checklists and the Mac users, because unclaimed work is how you come to own a system. Automate one repeat ticket type, measure what it saved, and write that down; it is the most persuasive single line in an internal promotion conversation. Then pick a branch and a certificate that match it: MD-102 and SC-300 for identity and Microsoft 365, Jamf or Intune depth for endpoint, Network+ then CCNA for networking, Security+ for a SOC analyst move, AZ-104 or the AWS equivalent for cloud. Ask whether your employer reimburses exams before you pay for any of them, because many do and almost nobody asks.
Put this on a resume in about a minute
Paste your history once and point it at the Help Desk Technician posting you are looking at. No account, no card.
Build my resume free More roles