The small-team support desk: a complete guide

Most support advice is written for teams with a support manager, a knowledge base owner and a QA process. If your team is three people and one of them is the founder, that advice is a list of things you will not do.

This guide is the other list: what a small team actually needs, in the order the decisions arrive, with the trade-off stated rather than hidden. Every section links to a longer piece if you want the detail.

1. Start with one channel, not seven

The temptation is to connect everything on day one. The result is an inbox nobody trusts, because nobody has proved that one channel works end to end.

Connect the channel your customers already use. Answer one real conversation through it. Only then add the next. Each step should end with a working desk rather than a half-configured one.

Free help desk setup, step by step

2. One inbox, or the team answers twice

A shared Gmail account works for one person and breaks at three. Everyone sees everything, nobody owns anything, and "who replied?" is answered by scrolling.

What is needed is not a bigger tool but three properties: assignment, status, and history attached to the customer. Website chat, email, WhatsApp, Telegram, WeChat, LINE and Messenger should all land in that same queue — otherwise you have built the seams you were trying to remove.

How a shared inbox works · WhatsApp · LINE · WeChat

3. Triage needs three primitives, not a workflow builder

Small teams do not lack good intentions; they lack a fast way to agree what to do next. Tags say what a conversation is about, priority says how fast it needs an answer, and views turn those two into a worklist.

Keep the tag list to five or six. If everything is high priority, nothing is. If the team still scrolls the whole inbox after two weeks, the taxonomy is wrong, not the team.

Tags, priorities and views

4. Saved replies should save structure, not sentences

A saved reply that reads like paragraph four of a script costs the trust it was meant to save. Save the structure — "here is what happens next, and here is when you will hear from us" — and leave a slot for the customer's specific situation.

Do not save answers to questions you have not answered. A saved reply that closes a ticket without resolving it teaches customers to distrust the channel.

Canned responses that still sound human

5. Where the data lives is an architectural decision

Cloud monitoring products upload records to the vendor; on-premise keeps them inside your network. The same question applies to a support desk, because it holds customer conversations, contact records and invoices.

Cloud is usually right for a team without infrastructure. Self-hosted is right when the data question is decisive — and it should be a supported option, not a special project. Before the first container: decide the host, test a restore, terminate TLS, and know how you will apply the next upgrade.

Cloud vs self-hosted checklist · self-hosting with Docker · deployment options

6. Price the work, not the seats

Per-seat AI pricing punishes the team that delegates work to the assistant: the better the automation, the more you pay for people. Per-action pricing tracks the work instead.

Whichever model, ask what counts as an action, whether a human can take over mid-thread, and whether the reply is visible in the conversation rather than in a log nobody reads.

What AI support costs per month · published pricing

7. The quote should become the invoice without re-keying

In most small companies the desk, the CRM and the invoice are three tools and a person moving data between them. Every manual transfer is a chance to lose the discount that was agreed, or the line item that never reached the invoice.

When the conversation creates the customer record, and the quote is built from the same line items the invoice uses, that whole class of error disappears.

Free CRM setup · CRM and invoicing

8. Migrate in order, verify the first bucket

Sort the old data into must-keep, nice-to-have and can-go. Migrate in that order. Then open ten random conversations in the new tool and compare them with the old one; if the history is wrong for ten, it is wrong for all of them.

Migrating a help desk

9. Week one is for proving the pipe, not for dashboards

Are conversations reaching the inbox? Does every conversation have an owner? Which three questions repeat? Those are the only questions worth answering in week one. First-response time on eleven conversations is noise, and CSAT with a handful of responses is decoration.

What a free ticketing system should include

Where Onsites AI fits

Onsites AI is a support desk with CRM and invoicing built in, offered cloud or self-hosted. Website chat, email, Telegram, WeChat, LINE, WhatsApp and Messenger land in one shared inbox; AI replies sit inside the conversation and any agent can take over; and the first three seats are free permanently, with AI billed per action from about $0.03.

Product overview · how it compares · FAQ · get started

Every guide, in reading order

Written by the Onsites AI team. Free for the first 3 seats. onsitesai.com

Comments

Popular posts from this blog

Handling refunds without a second system

A support policy that fits on one page