LiveChat Alternative for Ecommerce: When Per-Agent Pricing Stops Making Sense

A store that ships physical goods has a support queue with a predictable shape. Where is my order. Can I change the shipping address. Does this ship to Canada. How do I return it. LiveChat handles every one of those well, on one condition: somebody has to be signed in.

The friction that sends stores looking for a LiveChat alternative is what happens during the hours when nobody is, and what it costs to shrink those hours. LiveChat bills per person. As of publication, its pricing page lists Starter at $19/month per person billed annually, Team at $49, and Business at $79, with monthly billing at $25, $59, and $89 respectively. Covering more of the clock means more people on the account, so the bill tracks your coverage rather than your conversation volume.

What LiveChat is genuinely good at

LiveChat has been doing one thing for a long time and does it better than most tools that bolted chat on later. Its own pricing page claims 35,000+ companies as customers, and the product shows that maturity in unglamorous places.

The widget is the polished part. Themes, dark mode, a live editor with real-time preview, right-to-left support for Arabic and Hebrew, and widget translations across 48 languages. Multiple website support is included on every tier at no extra charge, so one account can cover a main store and a landing-page domain.

The agent-side tooling is where the years show. Smart routing distributes chats between available agents; manual routing lets a team queue everything and pick chats up by hand; routing rules can fire on URL, so traffic from /pricing goes to sales and /help goes to support; geolocation rules route visitors to agent groups that speak their language. Chats transfer between agents, and any agent can take over a conversation from a bot or a teammate. Canned responses come with private variants, event-based suggestions, and automated tags. Internal notes let the team talk behind the transcript.

For a team large enough to have a schedule, Business ($79/month per person, annual) adds a work scheduler for shift handoffs, staffing prediction that estimates how many agents you need, agent performance and agent activity reports, chat duration and chat response reports, and on-demand reporting exports. That is a workforce management layer, and it exists because LiveChat's core customer is a support department with a rota.

LiveChat also ships native Shopify, WooCommerce, and BigCommerce integrations on all four tiers, plus goals and a sales tracker that attributes chats to revenue. Facebook Messenger is native on every plan. SMS, WhatsApp, Instagram, X, and email arrive through the marketplace, and Apple Messages for Business is included from Business up.

None of this is a warm-up before the criticism. If you are staffing a chat queue with humans and want them to be faster and better managed, LiveChat is a reasonable place to spend money.

The seat math

The problem starts with what a seat costs and when you need one.

LiveChat's own FAQ is direct about the billing unit: "we charge a fee for every user account created, regardless of whether the user is logged in or not." A part-timer who covers Saturday mornings costs the same as your full-time lead.

Starter, at $19/month, allows one user. That is the tier a solo store owner can justify, and it is also the tier that cannot add a second person. Adding weekend coverage means moving the whole account to Team at $49/month per person, which is $98/month for two people. The first hour of weekend coverage takes the bill from $19 to $98.

LiveChat's FAQ works a larger example itself: for 10 agents and one chatbot, "you'll pay $52 for one chatbot, and the cost for each agent starts at $41," totalling from $462 per month. A 15% discount is available for a 12-month commitment. Enterprise pricing is not published; the page routes to a sales call. The full tier-by-tier breakdown is in our LiveChat pricing rundown.

Per-seat pricing is honest and it is standard. It just means the cost of answering a customer at 2am is the cost of paying somebody to be awake at 2am.

The AI is a separate product

LiveChat's plans do include AI, branded Text Intelligence and included on every tier. Copilot answers the agent's questions about LiveChat itself. Reply suggestions pull from imported knowledge sources, three on Starter and ten on Team and Business. There is chat summary, text enhancement, tag suggestions, Insights for spotting popular questions, and sentiment analysis in early access.

Read that list again with one filter: every item assists a human who is already in the conversation. Reply suggestions suggest replies to somebody. Chat summary summarizes for somebody. None of it covers the shift nobody is working.

The product that answers without a human is ChatBot.com, a different Text product with its own subscription. LiveChat's pricing page promotes it under the headline "Solve up to 80% of cases with AI chatbots" and lists it starting at $52/month billed annually. In the feature comparison table, "AI chatbots" appears under integrations and add-ons with that same figure attached rather than as a plan inclusion.

The same pattern applies to the rest of the suite. Email ticketing is HelpDesk.com, starting at $29/month. A customer-facing help center is KnowledgeBase, starting at $49/month. Private API usage begins with a free $25 credit and is billed separately after that.

So a store that wants storefront chat, AI resolution, email ticketing, and a help center is buying four subscriptions from the same vendor. LiveChat does not publish a bundled price for that combination, so the honest version is that the total depends on which products you take and at what tier. What is published is that each starts at its own number, and those numbers move independently.

One more detail worth knowing before you evaluate ChatBot.com on its own: LiveChat's FAQ states it "is a standalone AI. It doesn't rely on third-party providers like OpenAI, Google Bard, or Bing AI." That is a real architectural choice with real tradeoffs in both directions, and it is worth testing against your own transcripts rather than assuming either way.

What an ecommerce queue actually requires

For most stores the deciding factor has nothing to do with price.

"Where is my order" cannot be answered by a faster typist or by a chatbot trained on your policy pages. A bot with your shipping policy in its knowledge base can recite that orders ship in two business days. It cannot tell a customer where order #10482 is unless it can read the order record.

Returns are the same. The correct answer is either "yes, here is your label" or "no, that shipped 47 days ago and the window is 30," and both require the order date checked against the policy. Cancellations require knowing whether the order has been fulfilled yet.

LiveChat's ecommerce integrations put that store context into the agent's sidebar, which is exactly right for a human-staffed queue. The context is delivered to a person. If the person is asleep, the context sits there.

For most small stores, order status and returns are the majority of the queue by volume, and they are also the most mechanical. Handing that share to something with direct access to the store data is a different move than making your agents 20% faster at typing.

Who should stay on LiveChat

Stay if you have three or more people who will keep answering chats and you want them managed well. The work scheduler, staffing prediction, agent performance reporting, and group routing exist for that team, and no AI-first tool matches that layer.

Stay if a meaningful share of your conversations already happen on Facebook Messenger, WhatsApp, Instagram, X, SMS, or Apple Messages. Consolidating those into one agent inbox is a job LiveChat's marketplace does.

Stay if you need Enterprise controls. Single sign-on, HIPAA compliance, a white-label widget, SLAs, dedicated onboarding, and purchase-order billing are all listed on the Enterprise tier.

Stay if your queue is consultative rather than transactional. If every conversation is genuinely different, there is less for an AI agent to absorb, and the right investment is a better-equipped human.

Where Ernest fits, and where it does not

Ernest is an AI support agent built for Shopify and WooCommerce stores. It connects to the store, imports policies and the product catalog automatically, scrapes the rest of the site for FAQs and product details, and then answers on the storefront widget and by email, 24/7, in the customer's language.

The order questions are the point. Ernest looks up live order status, searches the live catalog for product and sizing questions, and can process cancellations and returns. Each of those actions is configurable per merchant as fully automatic or one-click approval. Refunds are always approval-gated, with no setting that changes it.

Before going live, a store can replay its own past tickets against Ernest in simulation and read how it would have answered. After going live, conversation intelligence surfaces the questions that keep coming up without a good answer.

Pricing is volume-based with no seat count attached: a free plan with 100 conversations that never expire (a real free plan, not a trial), then $49/month for 500 conversations, $149/month for 2,000, and $299/month for 10,000. AI is included at every tier, with no separate AI product to subscribe to.

The comparison that matters is a unit comparison, not a number comparison. LiveChat's Team tier is $49/month per person billed annually; Ernest's Starter tier is $49/month for 500 conversations regardless of how many people log in. Those two $49s buy different things, and picking between them is really picking whether you are buying coverage hours or conversation volume.

The limits, stated plainly:

• Two channels only. Storefront widget and email, with escalation to a human by email. No phone, SMS, WhatsApp, Instagram, or Messenger. LiveChat covers considerably more surface here.
• No agent workforce tooling. No work scheduler, no staffing prediction, no agent performance reports, no routing rules between agent groups. If you are running a rota, Ernest gives you nothing to run it with.
• No helpdesk, CRM, or returns-portal integrations, and no white-label or agency reseller option. LiveChat's marketplace exists precisely for that kind of assembly.
• Refunds always require a merchant click. If you want fully autonomous refunds, Ernest will not do it.

How to decide

Two questions settle this without trialling anything.

How many hours a week is nobody signed in? Multiply that by your chat volume during those hours. That number is what an AI agent is being hired to cover, and it is also the number that tells you whether adding a seat is cheaper. Our staffing calculator will do the coverage arithmetic if you want it precise.

What share of your queue is order status, returns, and cancellations? Pull last month's transcripts and count. Above roughly half, the bottleneck is access to store data, and a faster human does not fix it. Below that, LiveChat's agent tooling is aimed at your real problem.

If the answer points toward AI resolution, the practical next step is checking accuracy against your own history rather than a vendor demo. Shopify stores can install Ernest from the App Store and run the simulation on real past tickets first.

For a wider view of the field, our roundup of customer service chatbots covers the options beyond these two, and the Tidio alternative comparison walks through a tool with a similar chat-plus-separate-AI billing structure. If you want the pre-purchase side rather than post-purchase support, what an AI sales agent for Shopify covers goes into that separately.