How to Choose Live Chat Software: A 20-Point Checklist

Knowing how to choose live chat software comes down to matching a tool to the way your team actually works, not to the longest feature list. Most products can show a chat bubble and route messages to an inbox. The differences show up in the details: how fast operators get context, what the AI really costs, whether the mobile app is usable, and where your data lives. This 20-point checklist groups those details so you can compare tools side by side and run a short, honest trial.

Start with your use case: support, sales or both

Before you look at features, write down three things:

• Who will answer chats? One founder between meetings, a few people part-time, or a dedicated team.
• What will people ask? Support questions, pre-sale questions, or both.
• When? Business hours only, or around the clock with an AI assistant covering the gaps.

A two-person team needs speed, good notifications and low cost. A growing support desk needs assignment, statuses, analytics and data access. Keep that in mind as you go through the list — not every point matters equally for everyone. If you're still deciding between a bot and people, read chatbot vs live chat first.

Widget: install, look, mobile, attachments, performance

1. Install effort. A single script tag should be enough, plus ready-made plugins or guides for your platform (WordPress, Shopify, Webflow, Wix and so on).
2. Doesn't break your site. Look for isolation of the widget's styles (for example, Shadow DOM) so your theme and the chat don't interfere with each other.
3. Customization. Colors, light/dark theme, position, launcher size and greeting text — enough to look like part of your site.
4. Mobile experience. Test on a real phone: the window shouldn't cover the whole page awkwardly or fight with your sticky buttons.
5. Attachments. Customers should be able to send screenshots, photos and documents, with a sensible size limit.
6. Performance and reliability. The script should load asynchronously, and messages typed during a brief connection drop shouldn't be lost.

Operator tools: inbox, assignment, statuses, visitor context

1. One clear inbox with tabs or filters for unassigned, mine, and resolved conversations.
2. Assignment and statuses — open, snoozed, closed — so nothing sits unanswered in a shared queue.
3. Visitor context. Current page, device, approximate location, local time, referrer and, for signed-in users, name and email passed from your site.
4. Conversation tools. Quoted replies, editing sent messages, read status — small things that save confusion.

AI assistant: models, cost, handoff, control

1. Handoff to a human. The bot should transfer to an operator cleanly, with the full history, and the customer should know it's happening.
2. Control over behavior. You should be able to write the instruction, set rules for specific topics, and decide what the bot may do (hand off, send the visitor to a page, close the conversation).
3. Model choice and transparent cost. Ask which models are available and how AI is billed: per resolution, per conversation, bundled in a plan, or per token on your own key. Estimate a realistic monthly cost from your chat volume.
4. After-hours behavior. Outside working hours, the bot should tell visitors when people are back instead of pretending someone will answer now.

Apps and notifications: desktop and mobile

1. Notifications that actually reach you. Browser push, desktop apps with native notifications and tray icons, and a mobile app. For small teams, this matters more than almost any other feature: a chat nobody notices is a lost chat.
2. Usable mobile app. Not just alerts — operators should be able to read the context and reply comfortably from a phone.

Analytics and data access

1. Core metrics out of the box. Response rate, first response time, wait time after bot handoff, ratings, per-operator activity, and comparison with the previous period. Our live chat metrics guide explains which ones matter.
2. Getting your data out. Can you export transcripts or connect analysis tools? Some services offer an API or a read-only integration for AI tools; ask what's included.

Pricing: seats, add-ons, AI costs

1. Total cost, not headline price. Add up per-seat fees as the team grows, paid add-ons (branding removal, analytics, apps), and AI usage. Compare the cost for your team in a year, not today. For options with no seat fees, see our roundup of free live chat software.

Data storage and privacy questions to ask

1. Where data is stored and who processes it. Ask every vendor:

• In which country are conversations and visitor data stored?
• What visitor data is collected (IP addresses, location, device) and for how long?
• Which third parties process data — for example, the AI model provider?
• How do you delete a conversation or a whole account?

Requirements differ by country and industry, so check data residency against your obligations and consult a lawyer if you handle sensitive data. Don't rely on a sales page alone.

How to run a one-week trial

A trial only helps if you use the tool on real traffic. A simple plan:

1. Day 1. Install the widget on your site, customize it, invite everyone who will answer chats, and install the desktop and mobile apps.
2. Day 2. Send test chats from a phone and a private browser window. Check notifications on every device, attachments, and the visitor context operators see.
3. Day 3. If you plan to use AI, write a short instruction, test it with your 20 most common questions, and check the handoff.
4. Days 4–6. Run it on live traffic. Ask operators to note anything that slows them down.
5. Day 7. Review: response times, missed chats, ratings, operator feedback and projected cost. Decide with numbers, not impressions.

If two tools are close, the one your operators prefer using every day usually wins.

Frequently asked questions

What features should live chat software have?

At minimum: an easy-to-install widget, a shared inbox with assignment and statuses, visitor context, reliable notifications on desktop and mobile, ratings after conversations, and basic analytics. An AI assistant with clean handoff to a human is increasingly expected.

How much should live chat software cost a small business?

It varies widely. Many tools charge per operator per month, with extra fees for AI or add-ons; others have no seat fees. Calculate the total for your team size and chat volume over a year rather than comparing headline prices.

Do I need a mobile app for live chat?

For a small team, almost certainly. If people answer chats between other tasks, a mobile app with push notifications is often the difference between a fast reply and a missed customer.

How long should I trial a live chat tool?

A week on real traffic is usually enough to see response times, notification reliability and operator feedback. Shorter trials tend to test only the setup, not daily use.

Get started

ChatDeck covers this checklist with a one-line install, an operator inbox on web, iOS and desktop, live visitor context, analytics and an AI assistant on your own API key via the GPTunneL gateway, with no per-operator fees or paid add-ons. Note that its servers are in Russia, so check that against your data requirements. Create a free workspace for your trial, and compare options in our guide to free live chat software. To see the chat from both sides before installing, try the Try the chat before installing tutorial.