Help Desk vs Service Desk: How to Choose for Your Team

A help desk answers the questions customers send and fixes what they report. A service desk adds requests, approvals and change control across teams. A small software team supporting its own customers needs the first. Priced for 3 agents on 9 Nov 2026, the ITSM plans in our table run from $0 to $297 a month and the customer help desks from $49 to $75 billed annually, so entry prices are close and the process you must maintain is the real difference.

Key takeaways

• Help desk means break-fix and questions. Service desk means a single point of contact tied to ITIL incident, request and change processes.
• Most “help desk vs service desk” pages are written for internal IT. A team answering customers about its own product is doing help desk work with a few service-desk habits.
• Entry prices are close for 3 agents: Jira Service Management is free, and Freshservice and Zendesk both start at $57 billed annually. The ITIL tier is where cost jumps: Freshservice’s plan with a service catalog and SLA management is $147.
• Add the fourth agent and the numbers change. If $21.50 is the per-agent monthly rate, Jira Service Management goes from $0 to $86.
• Buy a capability when someone can name the work it will handle and who maintains it.

Help desk vs service desk at a glance

A help desk fixes reported problems and answers individual questions. A service desk is the single point of contact that also coordinates service requests, escalations and user communication across the service provider.

The historical distinction matters. HDI’s history of help desks describes a support function built around operational, break-fix work. Modern help desks can include automation, knowledge bases and analytics without changing that basic purpose.

HDI’s guide to the help desk-to-service desk transition explains the broader service desk role: handling incidents and requests while connecting users to service-management processes. The difference is operating scope, not whether the software includes chat.

Help desk and service desk side by side

What the differences mean for a founder

Core purpose and typical work tell you whether a shared inbox is enough. A customer asking how to configure a Shopify app usually needs an accurate answer, not an approval workflow.

Process scope and framework origin tell you how much coordination you’re buying. A service desk can keep a request moving between support, security and engineering, but somebody must maintain those workflows.

Metrics and team structure tell you what managers must own. A fast first reply doesn’t prove that a request reached the right specialist or that customers received useful outage updates.

Practical rule: A service desk isn’t a reward for growing. It’s a commitment to managing work across boundaries.

For a small software team, conversational product support is usually help desk work. Adding a named engineering escalation owner doesn’t require adopting an entire IT service management, or ITSM, suite.

Run this test today: questions about how something works or why it broke are primarily help desk work. Requests for controlled access, new services, planned changes or coordinated outage updates need service-desk practices.

What does the choice cost a 3-person team each month?

For 3 agents, a help desk and an ITSM service desk can cost about the same. Jira Service Management is free for 3 agents. Freshservice starts at $57 and Zendesk Support Team at $57. The bill splits when you need the ITIL parts: Freshservice’s tier with a service catalog and SLA management is $147 for 3 agents.

We priced the same 3-agent team on each vendor’s public page on 9 Nov 2026. These are list prices, with no AI usage, add-ons or discounts.

Monthly bill for 3 agents, list prices checked 9 Nov 2026

Source: Convot analysis of the vendors’ public pricing pages, 9 Nov 2026. Atlassian shows $21.50 per agent a month with a monthly and annual billing toggle and says annual saves up to 17%; we did not confirm which basis the displayed figure uses.

Three things stand out.

Free is real at 3 agents. If you only need a portal and queues, Jira Service Management costs nothing for three people. Convot’s Community plan is also free, for teams under $1,000 MRR, with 1 app and 3 seats.

The ITIL tier is the expensive step. Within Freshservice, the move from Starter to Growth to add a service catalog and SLA management takes 3 agents from $57 to $147. The top Freshservice plan is $297. Whether that is worth it depends on whether you have requests to catalogue and contractual targets to time, as the sections below describe.

The fourth agent matters most on the free plan. Here is the same team growing, at the same list prices.

Monthly bill as the team grows

Source: Convot analysis of vendor pricing pages, 9 Nov 2026. Convot figures from our pricing page: Starter includes 3 seats, extra seats are $10 each.

At entry level the price gap between a service desk and a help desk is small; at the ITIL tier it is not. The difference that stays either way is operational. A service catalog, approval flows and change records need an owner. If nobody on a 3-person team has that job, the plan’s extra modules are shelfware you still pay for.

Convot’s pricing is by apps and includes seats, so a team with several products pays one bill. Look at the shared inbox to see how conversations from every app land in one place. If you are still defining what the desk is for, start with what a help desk is for a SaaS team, which includes a 30-day split of what AI answers and hands off.

Incidents, requests and who owns each

An incident is an unplanned interruption or reduction in service quality. A service request asks for an agreed service, information or access without reporting a disruption, as distinguished in the ITIL Service Operation guidance.

A checkout extension failing during purchases is an incident. A merchant asking for a sandbox account is a request. Both can arrive through chat, but they shouldn’t inherit the same priority or completion criteria.

Separate technical ownership from communication ownership

The service desk responsibilities described in Micro Focus’s ITIL 2011 best-practice guide provide a useful operating sequence:

1. Log, categorize and prioritize every contact. Record the affected app, symptom and business impact. A checkout failure needs different handling from a question about an invoice.
2. Run first-line diagnosis. Ask for the error message, reproduction steps and affected environment. Don’t send engineering a screenshot without enough context to investigate.
3. Resolve incidents and requests where possible. Support might explain a configuration fix or fulfill an already authorized sandbox request. Permission to answer isn’t permission to grant privileged access.
4. Escalate unresolved cases within agreed time limits. Route the checkout failure to the engineer responsible for that integration. Include the investigation already completed and a clear next action.
5. Keep users informed through resolution. Give the merchant a useful update even when engineering hasn’t found the cause. Don’t make customers repeat their story because the technical owner changed.

A narrow help desk may implement the intake, diagnosis and resolution steps well while leaving escalation and updates informal. That’s not an unavoidable limitation of help desk software. It’s an ownership gap.

Give each workflow a finish line

An incident finishes when service is restored and the customer understands the outcome. A request finishes when the agreed action is completed, rejected with a reason or returned for missing information.

For recurring faults, keep the immediate restoration separate from the investigation into the underlying cause. Otherwise, support may hold a customer’s ticket open indefinitely while engineering schedules a permanent fix.

Count the disruptions and requests in your queue separately. Then mark which category regularly leaves support’s control. That split tells you where a simple assignment rule is enough and where a defined fulfillment workflow earns its maintenance cost.

Feature map for each desk and what SaaS teams need

A feature list tells you what a tool includes. The capability map tells you what your team must operate.

Treat these as typical expectations, not guarantees attached to a product name. A help desk can offer advanced routing, while a service desk still needs people to configure and maintain its processes.

Feature levels by desk type and SaaS need

Avoid paying to maintain unused processes

Change and asset management cause unnecessary spending when founders confuse customer support with internal infrastructure management. Explaining an app setting doesn’t require a hardware inventory or a formal change advisory process.

Controlled production changes are different. If support can initiate a data migration, alter permissions or request a production configuration change, record authorization, execution and rollback ownership. You can introduce that discipline without buying every module in an enterprise suite.

Formal SLA tracking deserves similar scrutiny. Contractual response or restoration commitments need dependable timers, calendars and escalation rules. Internal service targets can start with a documented policy and regular review, although customers still deserve clear expectations.

Buy a capability when someone can name the work it will handle and the person responsible for maintaining it.

Reuse knowledge before adding automation

A shared answer source is more useful than separate collections of chat scripts and help articles. Write the setup instructions once, maintain the canonical article and reuse it in replies.

Our product includes a self-serve help center for each app, a changelog and roadmap, a status page for uptime and incidents, and scheduling for booking calls in chat. A lightweight help desk can therefore cover several customer-facing needs without becoming a full service-management suite.

Use the help desk software features checklist to separate essentials from attractive extras. Mark each capability needed, optional or not needed, then review last month’s tickets and identify which ones would have used it.

Count actual work, not hypothetical future use. An unused workflow still costs configuration time, training and attention.

Use cases by team size and product count

Team size matters less than the work people must coordinate. A solo founder can need approval controls, while a larger customer support team can operate effectively with a help desk.

Best-fit desk by team profile

These profiles are decision examples, not staffing thresholds. Intercom’s service desk explanation places the distinction in scope: a support queue handles reported issues, while the broader service desk coordinates requests, service commitments and ownership.

Multi-app support needs context before ceremony

Say a merchant uses several apps from your studio. Separate inboxes can force agents to search for earlier conversations before they can understand the current question.

Our shared inbox places every app’s conversations together, labels each conversation by app and keeps one customer record across apps. Saved views, notes, @mentions, assignment and first-response analytics help teammates divide the work without splitting the customer history.

For Shopify app developers, the merchant’s plan, MRR and past billing appear beside the conversation, pulled from the Shopify Partner API. That context helps distinguish a billing question from a setup issue and identify the relevant subscription without sending the merchant through another explanation.

Product identity still needs its own label. Revenue context doesn’t replace knowing which app, version or workflow the customer means.

Add only the service-desk habits you use

Pricing by app count can avoid separate help desk subscriptions for each product. Our model includes seats within each tier on one bill, with additional capacity available when the team exceeds the included allowance.

The operational trade-off is shared governance. Decide who maintains app labels, who owns cross-app customers and who covers conversations when the usual agent is unavailable.

• Keep structured intake. Collect the app and the information needed to investigate.
• Keep explicit escalation. Name the receiving team and the expected next action.
• Drop unused ceremony. Don’t require agents to complete internal IT fields that have no bearing on an external customer’s problem.

ITIL terminology is useful when it clarifies responsibility. When agents spend more time interpreting categories than helping merchants, simplify the workflow.

Cost, AI resolution and pricing tradeoffs

AI resolution volume changes the budget faster than the help desk label does. Convot’s Cove AI is $0.20 per resolved conversation and Intercom’s Fin starts from $0.99 per outcome, so at 1,000 AI-resolved conversations a month the AI line alone is $200 on Cove AI against about $990 at Fin’s entry rate, before subscriptions, add-ons or human work.

Confirm current rates and tier prices on each vendor’s pricing page before committing a budget.

Model the successful month

Per-resolution billing grows when the AI resolves more conversations. A quiet pilot can look cheap while the AI line item grows sharply in production.

Use this model:

Monthly support cost = base subscription + app and seat add-ons + billable AI resolutions + human handling and administration.

Cost items and what to check before buying

Convot’s monthly tiers are Starter $49, Growth $79, Pro $149 and Scale $299. Growth includes 2 apps and 7 seats, with $20 per extra app and $10 per extra seat. Flat team pricing means a defined allowance. Count only teammates who can reply: support reps, founders and contractors. Teammates who only read or leave internal notes don’t add seats.

Resolution quality belongs in the cost model

Cove AI is $0.20 per resolved conversation. It answers setup and billing questions from each app’s own documentation and hands off to a human when it isn’t sure.

A lower unit price still pays for mistakes. Measure reopened conversations, corrections and human investigation beside the vendor’s resolution count.

The AI customer service agent guide is useful when separating answer automation from actions that change an account. Explaining a billing policy and changing billing permissions need different controls.

Budget rule: Compare the cost of a correctly resolved customer problem. A conversation marked resolved is only the first line item.

Fin can still be sensible for a team already operating in Intercom if changing platforms would disrupt established workflows. Include migration, training and administration in that decision rather than comparing AI rates alone.

For Shopify teams, merchant billing context can inform prioritization. It doesn’t predict cancellations, and it shouldn’t override the severity of a widespread incident.

Confirm your support scope before you switch

Support work needs service-desk practices when it involves controlled changes, approvals, cross-team ownership and recurring-issue analysis. Mostly customer questions, bug reports and account guidance usually call for a help desk with selected coordination features.

The distinction becomes useful when you apply it to real conversations rather than vendor demos.

Audit the queue before approving the purchase

1. Classify a sample of recent tickets. Use your last 100 as a practical audit sample, separating incidents, requests and ordinary product questions.
2. Mark approval and ownership boundaries. Identify work that support couldn’t complete without another team’s permission or action.
3. Find communication obligations. Flag conversations requiring ongoing status updates, not just an initial reply.
4. Test uncertainty handling. Give the AI an ambiguous question and confirm that it preserves context when handing off to a human.
5. Calculate the full bill. Use your expected app count, replying teammates and resolved-conversation volume, with the tables above as a starting point.

Measure first-contact resolution separately for incidents and requests. HDI’s support metrics white paper quotes the 2013 HDI Support Center Practices and Salary Report: an industry average first call resolution of 68.8% for incidents and 68% for service requests, measured on phone contacts only.

Treat those figures as dated context from phone support, not targets for chat and email. A queue of routine access requests isn’t directly comparable with a queue of complex integration failures.

Record the contact channel, classification, first owner, handoffs and resolution timestamp. Pair resolution rates with reopen rates, escalation rates and customer effort so fast closures don’t conceal unfinished work.

Choose the operating model honestly

Disclosure: Convot makes a competing help desk product. Built by a solo founder, it fits small software teams, particularly Shopify app developers, that want shared support by app rather than a full internal ITSM implementation.

Our AI answers from help center documentation and account data, cites its sources and hands off when it isn’t sure. The Community plan is free for teams under $1,000 MRR, with 1 app, 3 seats and no Cove AI.

A seat means someone who can reply, including founders and contractors. Read-only collaborators and teammates who only leave internal notes don’t count.

A full ITSM suite is a fairer choice when change approvals and asset tracking drive daily work. For organizations weighing outsourced infrastructure support, RDE Technologies’ managed IT services in NYC provides a relevant service option to evaluate alongside internal staffing.

Run the audit, choose the smallest operating model that handles the work, and start on the free plan if its limits fit. Don’t migrate just to rename the queue.

How to move support without losing conversation history

The migration path depends on your current tool. Crisp imports in one click; Zendesk, Intercom and Freshdesk moves are free migrations done with our founder, not automatic importers.

Book that migration session before exporting or canceling anything. Help Scout has no self-serve importer, so talk to us before switching; for any other source, confirm the path first.

Move the history before redirecting customers

1. Confirm the source and migration method. Establish whether you’re using the Crisp importer or a founder-assisted move. The common mistake is assuming that every advertised migration follows the same process.
2. Inventory what must survive. List conversations, customer records, internal notes and help center articles, then confirm which fields can transfer. Don’t discover after cancellation that a required attachment or identifier wasn’t included.
3. Map categories to app labels and saved views. Decide how agents will find old conversations under the new structure. Reusing ambiguous labels can hide relevant history even when the records transferred correctly.
4. Validate a small sample. Inspect long threads, repeat customers and conversations containing internal notes. Check customer associations as well as message text, because readable history attached to the wrong record still creates support problems.
5. Switch the chat widget on one app first. Confirm routing, customer identification and agent access before changing the remaining apps. Switching every product together makes it harder to isolate a configuration mistake.
6. Retain a temporary read-only archive. Keep the old tool accessible for a few weeks where its contract and data-retention terms allow. Don’t let agents continue replying from both systems, or the conversation history will split again.

Treat documentation as part of the migration

Help center articles train Cove AI, so the same write-up supports customers in the help center and chat. Review obsolete instructions, broken references and product-specific billing rules before enabling AI answers.

Use the help desk migration checklist to assign owners for exports, label mapping, documentation review and cutover testing. Give each task an acceptance check rather than marking the entire move complete when the import finishes.

For a multi-app studio, plan each app’s switch separately while keeping shared inbox labels consistent. Name the person who can pause the cutover, and tell agents where to look when a historical thread is missing.

FAQ

Is a help desk the same as a service desk?

A help desk and a service desk differ in scope: a help desk fixes reported problems and answers questions. A service desk is the broader single point of contact that also handles requests, communication and links into ITIL processes such as change management.

Do we need a service desk if we only support customers?

A team that only supports customers rarely needs one. Questions, bug reports and account guidance are help desk work. Add service-desk habits such as named escalation owners and outage updates, and buy ITSM modules only when you have approvals or change control to run.

What does an ITSM tool cost for a 3-person team?

On 9 Nov 2026, Jira Service Management is free for up to 3 agents and $64.50 a month on Standard. Freshservice’s three plans are $57, $147 and $297, and Zendesk Support Team is $57 annual. Help Scout Standard is $75 annual.

Can a small team use Jira Service Management as its customer help desk?

Jira Service Management can serve as a customer help desk at small scale. The Free plan includes a customer portal and email, chat and widget channels for up to 3 agents. At the fourth agent it moves to the Standard price. The question is whether you want to configure request types and workflows yourself.

What is the free option for a small software team?

Jira Service Management Free covers 3 agents. Convot’s Community plan covers 1 app and 3 seats for teams under $1,000 MRR, without Cove AI. Check each vendor’s limits before you commit.

Convot offers small software teams a shared inbox across apps, customer-facing support tools and an AI agent that hands off when it isn’t sure. If your audit points to a help desk rather than full ITSM, see the shared inbox and the flat pricing, or visit Convot to review the free plan and plan a migration that preserves your conversation history.

Written by Tarang Agarwal, founder of Convot. Vendor prices checked on the vendors’ public pages on 9 Nov 2026.