The short version: for a small business, the features that matter are canned responses, smart routing to a real human, offline forms that don’t get ignored, and reporting you’ll open once a week. Chatbots, fancy triggers and AI add-ons are nice extras, not the reason live chat works or fails. Buy for staffing reality, not for the demo.
Why most feature checklists get this wrong
Every “best live chat software” article lists the same twenty features: proactive triggers, co-browsing, video chat, AI chatbots, unlimited agents, sentiment analysis. Most small businesses will use maybe five of those twenty, ever. I know because I’ve run live chat on my own sites for the better part of a decade, and I’ve watched clients buy the top-tier plan for features they clicked “on” once and never touched again.
The right question isn’t “what can this software do.” It’s “what will a two-to-five-person team use every single day for the next twelve months.” That’s a much shorter list.
Canned responses and saved replies (the boring feature that saves the most time)
This is the single highest-value feature and it’s the least talked about because it’s not exciting. Saved replies let you write your answer to “do you ship to Ireland” once and fire it in two clicks forever.
When I ran chat on my own consultancy site, I tracked this for three weeks in a spreadsheet, because that’s the kind of thing I do. Out of 23 chats, 14 were variations of five questions: pricing, availability, whether I work with US clients, whether I do one-off projects or retainers only, and whether I do the work myself or hand it to a team. Once I had five saved replies covering those, my average handling time dropped from just under four minutes per chat to under ninety seconds. That’s not a small efficiency gain, that’s the difference between chat being a pleasant extra and chat eating your afternoon.
If you’re comparing tools, ask specifically: can non-technical staff create and edit saved replies without calling support? Some tools bury this in an “admin” tier you’d need to pay extra for.
Routing rules that don’t dump everything on one exhausted person
If you’re a team of one, routing doesn’t matter much. If you’re three people or more, it matters enormously. You want rules that send billing questions to the person who knows your Stripe account, and send everything else to whoever’s marked as online.
The failure mode I’ve seen most often is the opposite problem: a tool with no routing at all, so every chat pings every agent’s phone at once, and after a week someone quietly mutes the notification. Then chats sit unanswered for twenty minutes while everyone assumes someone else picked it up. If you’re comparing options for an ecommerce store specifically, I’ve written a longer breakdown of how to choose live chat support software for ecommerce that goes into routing by product category and order value, which matters more there than in a service business.
Offline forms and missed-chat capture (most businesses lose leads here)
Here’s an uncomfortable truth nobody selling you the software wants to say out loud: the offline experience matters more than the live one for most small businesses, because most small businesses are not staffed to answer chat live for eight hours a day, five days a week. If your chat widget goes offline and just disappears, or shows a generic “we’re not here” with no way to leave a message, you have built a lead-loss machine that looks like customer service.
Test this before you buy: close the app, wait, and send yourself a chat as if you were a visitor at 9pm. Does it capture an email? Does it ask what they need? Does it reach you the next morning as a proper lead, or does it vanish into a dashboard nobody checks? I have seen businesses run chat for a year and never once look at their “missed conversations” tab. That tab often has more sales in it than the live chats do.
The chatbot question, answered honestly
AI chatbots are the feature every vendor pushes hardest right now because it lets them charge more per seat. For a small business, a basic bot that answers FAQs and books a call is useful. A bot dressed up to sound like a real person, without saying it’s a bot, is a trust problem waiting to happen.
I once tested a chatbot on a client’s site that was configured to use a person’s first name and a friendly headshot, with no disclosure it was automated. A visitor typed “I know you’re not real, just give me a straight answer,” and left the site thirty seconds later without converting. That one exchange told me more than any vendor case study: people forgive a bot that’s honest about being a bot far more than they forgive one pretending to be a human. If you’re weighing whether AI chat is worth the extra cost, the deciding factor should be whether it saves your team real hours on repeat questions, not whether it “feels human.”
Proactive triggers, used sparingly
Proactive chat, where a message pops up after someone’s been on a pricing page for 20 seconds, works when it’s specific and rare. It annoys people when it’s generic and constant. “Need help?” popping up on every page within five seconds is the fastest way to make visitors close your widget and never open it again. Set triggers on high-intent pages only: pricing, checkout, booking. This is closely related to the kind of intent signals used in online scheduling software, where the same “someone’s stalling on this page” behaviour usually means they’re one nudge away from booking a call.
Integrations that stop double data entry
Ask specifically whether the chat tool pushes a transcript and contact details into your CRM or email list automatically, or whether someone has to copy and paste it. Zapier bridges exist for almost everything, but a native integration with your CRM (HubSpot, Mailchimp, Pipedrive, whatever you use) saves a real weekly chore. If chat leads sit in the chat tool’s own dashboard and nowhere else, they get forgotten. This is the same principle that applies when choosing online booking software: the tool that “does everything” is worth less than the tool that talks to the two systems you already run your business on.
Reporting you’ll open, not reporting that looks good in a demo
Most chat dashboards show you things like “average handle time” and “customer satisfaction score” with a big colourful graph. Fine, but the report that changes decisions for a small business is much duller: how many chats came in, how many became leads or sales, and which three questions came up most often this week. If your tool can’t show you that in under a minute of looking, you’ll stop looking, and then you’re paying monthly for a tool you’re flying blind on.
Mobile notifications (the feature nobody demos)
Nobody shows you this in a sales call, but it’s the feature that decides whether chat gets answered. If you or your team are away from the desktop most of the day, which describes most small business owners, you need a mobile app that pushes a proper notification, not just a badge on an icon you have to remember to check. Test this on your own phone before you sign anything. Set the app up, walk away, have someone send a test chat, and time how long it takes for your phone to buzz.
The staffing truth vendors won’t tell you
Here’s the part that gets left out of almost every “best live chat software” list: the software is rarely the reason chat fails for a small business. The reason is that nobody’s assigned to answer it, and it turns into a guilty red notification badge that everyone ignores. I’ve seen this at three different small businesses I’ve consulted with over the years, all with perfectly good software and a chat widget quietly rotting on their site, unanswered for hours, actively damaging trust because a visitor can see the widget says “we typically reply in a few minutes” and then nobody does.
Before you buy anything, decide who owns chat, during which hours, and what happens outside those hours. If the honest answer is “nobody, most of the time,” you’re better off with a simple offline form and a note that says “we reply within one business day,” set clearly in the widget, than with a live chat tool that promises instant response and never delivers it. An unmet promise is worse than no promise at all.
What it costs
Rough current pricing so you’re not surprised: basic plans from tools like Tidio or Crisp start free or around $19 to $29 a month for a small team. Mid-tier tools like LiveChat or Zendesk’s chat product run closer to $19 to $59 per agent per month once you add routing, integrations and reporting. Intercom sits higher again, often $74 and up per seat once you add their AI features, and that’s where a lot of small businesses overpay for capacity they’ll never use. Before signing an annual contract, it’s worth running through a proper checklist, and I’ve laid out the exact questions to ask any vendor in how to research a new software tool before signing up, most of which apply directly here: what happens if you cancel mid-year, what’s locked behind an “enterprise” tier, and whether the free trial lets you test with real traffic or just a demo account.
Online chat itself isn’t new; it’s been part of how businesses talk to customers for well over two decades, evolving from basic web widgets into the routed, integrated, sometimes AI-assisted tools we have now (see Wikipedia’s overview of online chat for the history if you’re curious). What’s changed isn’t the core idea. It’s how many extra features get bolted on to justify a higher price, most of which a small business will never open twice.
A short checklist before you buy
- Can non-technical staff create and edit saved replies themselves?
- Does the offline form capture an email and reach a real inbox, tested by you personally?
- Is the mobile app notification instant and loud enough to notice mid-task?
- Does it push transcripts into the CRM you already use, without manual copy-paste?
- Can you see weekly chat volume and top questions in under a minute, without building a custom report?
- Have you decided, in writing, who answers chat and during which hours, before you install the widget?
Frequently asked questions
Do small businesses really need a live chat chatbot, or is a human enough?
A basic bot handling FAQs outside working hours is worth having, but it should say clearly it’s automated. For most small businesses under five staff, a human answering during set hours plus a simple offline form outperforms an unlabelled AI chatbot pretending to be a person.
What’s the single most important live chat feature for a small business?
Saved or canned replies. They cut handling time dramatically on repeat questions and are the feature teams use every day, unlike proactive triggers or video chat which get switched on once and forgotten.
How much should a small business expect to pay for live chat software?
Expect roughly $0 to $29 a month for a basic plan covering one or two agents, and $19 to $59 per agent per month for a mid-tier plan with routing and CRM integration. Higher-end tools like Intercom run $74 and up per seat once AI features are added, which is often more capacity than a small team needs.
What happens if I install live chat but can’t staff it consistently?
Then set expectations honestly in the widget, something like “we reply within one business day,” and rely on the offline form rather than promising instant live response. An unanswered chat that promised speed damages trust more than no chat widget at all.
Related reading: How to Use Live Chat for Lead Generation in a Small Business and What Outsourcing Live Chat Really Involves for Small Businesses.
For the bigger picture, see my full guide to productivity.
