Live Chat Software Pricing: What You Actually Pay in 2026

Ask five live chat vendors what their software costs and you get five different answers, because each one is solving a different math problem. One charges per agent, one charges by conversation volume, one bills for every conversation its AI closes, and several stack a subscription on top of usage fees. The billing unit changes from vendor to vendor, so the sticker prices are not comparable to begin with.

The question worth answering is what your team pays at your agent count, your conversation volume and your AI usage. So this guide sorts the four billing models you will meet, shows what the advertised number leaves out, puts current published figures from eight vendors side by side, and works through three team scenarios you can hold your own quote against.

Part 1: The four pricing models you will run into

Most live chat software pricing falls into four patterns. Plenty of platforms combine two of them, which is what makes estimating hard.

1 Per-seat pricing

A flat monthly fee for every agent with a login, usually with no cap on how many conversations they handle. LiveChat lists Starter at $19 per agent per month billed annually and Team at $49. The formula is monthly cost = paid seats × price per seat .

This is the easiest model to forecast and the least responsive to a traffic spike you absorb without hiring. The catch is feature gating: a cheap seat often sits below the tier that carries ticketing, routing, advanced reporting, SSO or AI assistance , so the plan you can actually use may not be the plan you priced.

2 Per-conversation or per-contact pricing

Volume is the meter, which makes the entry subscription look inexpensive and lets the bill climb with demand. Tidio lists Starter at about $24.17 a month billed annually with 100 billable conversations, and Growth at about $49.17 a month covering up to 2,000 billable conversations. Above that, Tidio moves customers to custom pricing. The formula is monthly cost = base plan + additional billable conversations .

What separates vendors here is the definition, not the rate. Two platforms can both advertise usage-based billing while counting a "conversation" or a "contact" in ways that differ by a factor of several — ask for the counting rule in writing before you model anything.

3 AI-resolution or AI-usage pricing

The newest meter charges for each conversation an AI agent closes without a human. Intercom's Fin is $0.99 per successful outcome on top of the seat fee, and Help Scout lists AI Answers at $0.75 per resolution. The formula is monthly cost = base subscription + paid AI resolutions × rate .

This is the layer buyers underestimate, because it scales with your automation succeeding. A plan that looks cheap at 10 AI resolutions a month is a different product at 2,000 — and the vendor's revenue rises exactly as your deflection programme starts to work.

4 Consumption-based or custom quote

Enterprise platforms often price around a broader usage metric or build the quote around your requirements — several channels, larger teams, SLAs, implementation services, custom security. Zendesk publishes plans for small and mid-sized teams while keeping Suite Enterprise behind a sales conversation. The formula is total cost = negotiated subscription + usage charges + add-ons + implementation + tax .

5 Hybrid plans are now the norm

Few vendors pick one model and stop. The common shape is a per-seat fee for humans, plus a per-outcome fee for AI, plus premium channels and reporting tiers layered on top. The practical effect is that "cheapest plan" and "cheapest for my team" are usually two different vendors. A three-agent desk with little automation does better under a pure seat model, where the volume meter simply does not exist. A team leaning hard on AI deflection does better where automation is priced in flat session blocks instead of per outcome.

Part 2: What the advertised price leaves out

The number on a pricing page is rarely the number on the invoice. These items open the gap.

Annual commitment comes first. Most vendors display a monthly-equivalent figure that applies only if you pay for twelve months up front, and the month-to-month rate is higher. Zendesk's Suite Team is $55 per agent per month on an annual commitment; the monthly-billing rate is higher, so confirm the exact figure for the term you intend to sign.

Overages come second. If a plan includes 100 billable conversations or a fixed pool of AI resolutions, the question is what happens at 101 — the per-unit rate above the allowance, and whether you can set a hard spending cap.

AI add-ons deserve their own line on the quote. Intercom charges per Fin outcome and Help Scout charges per AI resolution, while HubSpot runs AI on credits: Service Hub Professional includes 3,000 credits and its Customer Agent consumes 50 credits per resolved conversation, which works out to a specific and forecastable cost once the included pool is gone.

Then the items no headline number carries: WhatsApp and SMS add-ons priced per message, feature-gated tiers, integration and implementation fees, data retention limits, support-tier upgrades, and tax. HubSpot requires one-time onboarding of $1,500 on Professional and $3,500 on Enterprise — a real line item that most comparison roundups omit entirely. Many vendors also display prices excluding VAT or GST, calculated at checkout.

Part 3: Live chat software pricing comparison, eight vendors

Every figure below was read from the vendor's own pricing and plan-detail pages in 2026, in USD, on annual billing unless noted. These are examples of different pricing structures, not a ranking.

Part 4: Three worked cost scenarios

The fastest way to understand live chat software pricing is to hold a workload constant and run it through several meters. Assumptions are stated so you can substitute your own numbers.

1 Three agents, 1,000 conversations, website-first

A small team wants website chat, predictable billing and no paid AI usage. LiveChat Team at $49 per agent gives 3 × $49 = $147 a month , or $1,764 a year before tax. Zendesk Suite Team lands at 3 × $55 = $165 and includes a small automated-resolution allowance a team this size will rarely exhaust.

The instructive part is what does not move the bill. Both plans are seat-based, so the 1,000 monthly conversations create no separate charge at all. At this size, headcount is the only variable that matters.

2 Ten agents, 5,000 conversations, 1,000 resolved by AI

Now automation is real. Intercom Advanced is $85 per user, so seats cost 10 × $85 = $850, and Fin adds 1,000 × $0.99 = $990, for $1,840 a month . Help Scout Plus at $45 per user gives 10 × $45 = $450 in seats plus 1,000 × $0.75 = $750, for $1,200 a month .

The 5,000 conversations are not the pricing unit here — the 1,000 AI resolutions are. That single line is the difference between the two totals, which is why an automation forecast belongs in every pricing model you build.

3 Thirty agents, 20,000 conversations, multi-channel with ticketing

At 30 agents with 8,000 conversations resolved by AI, Intercom Expert seats run 30 × $132 = $3,960 and Fin adds 8,000 × $0.99 = $7,920, reaching $11,880 a month . Zendesk Suite Professional seats are 30 × $115 = $3,450, and the Copilot add-on at $50 per agent brings it to about $4,950 — before any AI-resolution charge, which Zendesk does not publish.

This is where per-outcome AI pricing overtakes everything else. Intercom's AI line alone is larger than most teams' entire software budget, and it grows in direct proportion to how well the automation performs. At this scale a flat conversation tier or a session-block model changes the shape of the bill completely — which is the comparison worth running before you renew.

Part 5: How to compare a quote before you sign

Once real quotes arrive, run each one through the same worksheet so you are comparing like with like. Fill every row, including the ones that feel obvious — the rows teams skip are the ones that reappear as invoice surprises .

1 Which variable changes the bill fastest

Under pure seat pricing it is headcount: ten more agents at $49 adds $490 a month and nothing else moves. Under AI-outcome pricing it is resolution count — another 100 Fin outcomes adds $99, with no hiring involved. Under conversation-based plans the risk is the tier edge , where crossing an allowance pushes you into a higher band or into custom pricing, so your growth forecast matters as much as this month. For what to automate first, see our guide to customer service automation .

Part 6: When free live chat is enough

A free or entry tier genuinely covers a narrow case: a single-agent site, low inbound volume, no automation requirement, and no need for long chat history. Crisp's free plan runs two seats with a 100 customer-profile ceiling and no AI credits. Help Scout's free plan covers 5 users, 1 inbox and 1 Docs site.

The upgrade trigger is operational, and it is usually one of these: longer history, advanced routing, more agents, AI automation, removing vendor branding, real reporting, security controls, ticketing depth, or additional channels. If more than two of those apply to you today, budget for a paid plan now ; migrating a live support desk mid-quarter costs more than the tier you avoided. When ticketing is the trigger, our guide to automated ticketing systems covers what that layer has to do before it earns the upgrade.

Part 7: When unlimited-agent pricing makes sense

Every model in Part 1 taxes something, and per-seat pricing taxes access. The cost of that shows up as rationing, not as a line on the invoice. Teams share a supervisor login, leave QA reviewers outside the tool, and decide the weekend volunteer is not worth another $49 a month — support-quality decisions made for a licensing reason.

Look down the Part 3 table and every vendor but one puts a price on people, or a ceiling on them: Intercom, Zendesk, LiveChat, Freshdesk Omni and Help Scout charge per seat or agent, Crisp caps seats per tier and charges $10 for each extra, and Tidio holds most plans to ten. Only a conversation-metered plan makes the next login free , which is why the people who gain most are the ones a seat model prices out — supervisors and QA who read far more than they reply, night and weekend cover, seasonal overflow, and specialists who drop in a few times a month.

EngageLab prices LiveDesk against conversation volume instead, in three published tiers: $99, $299 and $399 a month for up to 30K, 100K and 150K conversations / month , the top tier expandable, each including 100 free AI credits / month , with support agents unmetered and no per-seat line at any tier . Read the tier list before anchoring on the entry price — automated ticket management appears from the $299 Business plan up , so a team that needs case ownership is comparing $299, not $99.

Put that beside Part 4 and the crossover is arithmetic. At ten agents with heavy AI use, the $1,200 to $1,840 range in Scenario 2 sits well above a flat conversation tier at the same volume. It flips the other way when your team is small and your conversation count is high , because a conversation-capped tier keeps climbing while a three-seat bill does not.

Two things still need a sales conversation, and neither should be estimated: the AI credit rate above the included 100 a month, and which meter governs your contract — the pricing page summarises the product as MEP-scale payment while the plan table gates on conversations. Ask which one is written into the agreement. None of this makes any product cheaper by default; it makes the comparison possible on paper, which is more than most vendors on this list allow.

Part 8: Your final buying checklist

Before signing anything, collect the evidence that actually predicts the bill.

• Your monthly conversation count and its trend across the last two quarters , not a single month.
• A realistic AI automation rate, measured on your own tickets instead of taken from a vendor's demo.
• The channels you need on day one, listed separately from the ones you might add later.
• Your headcount plan for the full contract term, including part-time and seasonal staff who need access.
• The written rate above every included allowance, and whether a hard spending cap is available.

Then run a usage-based trial for one realistic operating cycle, tracking conversations, AI resolutions, active agents, channels and admin hours. The goal is to learn which activity drives the bill, and how that changes as you grow.

Frequently asked questions

How much does live chat software cost?

Published entry plans run from about $19 to $55 per agent per month on annual billing, and flat conversation-based tiers start around $99 a month. No single monthly figure describes the category, because the billing unit differs by vendor. A three-agent team on a seat plan can run under $150 a month, while a thirty-agent team using AI heavily can exceed $11,000 — the same software category, four orders of decision apart.

Can live chat software be free?

Yes, for a narrow case. Crisp offers two seats with a 100 customer-profile limit, and Help Scout offers 5 users with one inbox and one Docs site. Free tiers stop being viable once you need chat history beyond the retention window, routing rules, AI automation, removal of vendor branding, or reporting anyone will act on.

Why is annual billing cheaper than monthly?

Vendors discount the commitment, and the gap is larger than most buyers expect — commonly in the range of a fifth to a quarter of the price. The published "per month" figure on most pricing pages is the annual-commitment rate. Confirm the monthly rate separately if you need contract flexibility , and price that flexibility as the real cost it is.

What is an AI resolution, and how is it billed?

It is a conversation the AI closes without a human, but each vendor defines "closes" differently, and the definition is the part that costs money. Some count a confirmed resolution, some count a customer who simply stops replying. Get the vendor's exact definition in writing, then track deflection rate and satisfaction together — a rising deflection rate with falling satisfaction means the AI is ending conversations, not solving problems.

Which pricing model is cheapest?

Whichever one meters the thing you have least of. Few agents and high volume favours seats; many occasional users and moderate volume favours a flat conversation tier; heavy automation means comparing per-outcome rates against session blocks, where the gap at volume runs to thousands a month.