Skip to main content
EzyConn

Intercom Pricing Explained: What You Pay at 5, 10 and 25 Agents

The confusing part of Intercom pricing is not the price. It is that there are two meters running at the same time, and only one of them is on the plan page. Here is the arithmetic worked out at three team sizes, in plain numbers you can copy.

EzyConn EditorialThe EzyConn blog 6 min read Updated

The two meters, in one paragraph

Meter one is per seat. Intercom Essential is $39 per seat per month on monthly billing. Meter two is Fin, the AI agent, at $0.99 per resolution. Your bill is the first multiplied by your headcount, plus the second multiplied by how often the AI succeeds. Both figures were checked in July 2026, so confirm current published rates before you budget on them.

Meter one: seats

Per-seat pricing means you pay for every person who can log in and answer, whether they handle one chat that month or a hundred. At $39 per seat on Essential, the arithmetic is straightforward: five seats is $195, ten seats is $390, twenty-five seats is $975. Every month. Before the AI does anything.

The hidden cost here is not the price of a full-time agent. It is the price of a part-time one. A weekend volunteer, a night-shift contractor, or a developer who answers three technical questions a month all cost the same as your busiest agent. This is why teams start sharing logins, which saves money and destroys your reporting at the same time.

Meter two: AI resolutions

Fin is billed at $0.99 per resolution. A resolution is a conversation the AI finished on its own, without a person taking over. It is not a message and not a chat session. One long back-and-forth the AI closes by itself is one resolution and one charge.

That is a fairer unit than charging per message, and it is worth saying so. You are paying for work done, not for traffic. The problem is not the unit. The problem is what the unit does to your budget as the automation improves, which is the next section.

The bill at three team sizes

To make the comparison fair, every row below assumes the same thing: 50 AI resolutions per agent per month. That is a middling assumption. Substitute your own number and the shape of the answer does not change.

TeamSeats at $39AI resolutionsFin at $0.99Monthly total
5 agents$195250$247.50$442.50
10 agents$390500$495.00$885.00
25 agents$9751,250$1,237.50$2,212.50

Two things jump out. First, the AI line is bigger than the seat line at every size, which is the opposite of what most people expect when they budget. Second, a twenty-five person team is looking at roughly $26,550 a year on these assumptions. That is a hire.

The part that catches people: success costs money

Hold the team at ten seats and change only one variable, how often the AI closes a conversation by itself.

Scenario, 10 seatsSeatsFinTotal
AI closes 250 a month$390$247.50$637.50
AI closes 500 a month$390$495.00$885.00
AI closes 1,000 a month$390$990.00$1,380.00
AI closes 2,000 a month$390$1,980.00$2,370.00

Same team. Same plan. The bill nearly quadruples because the AI got better at its job. Nothing went wrong here. You wrote better help articles, the AI answered more questions without a human, and the invoice grew to match.

The awkward bit is where those two effects land. The saving shows up in your staffing budget. The cost shows up in your software budget. Those are usually two different line items with two different owners, and only one of them gets questioned in the quarterly review.

How to estimate your own number

Three sums, not one. Work out the bill for a quiet month, a normal month, and your worst month. Seasonal businesses get caught here, because a per-resolution meter charges most in exactly the month your revenue is already stretched.

Then run a fourth sum with the AI volume doubled. Not as a worst case, but as the success case. If that number would trigger an uncomfortable conversation with finance, you have found the real constraint on how good you are allowed to let your automation get.

When these numbers are fine

Plenty of teams look at $885 a month and shrug, correctly. If one closed deal is worth more than a year of the subscription, the software cost is noise. If your AI volume is genuinely low, the second meter barely registers. And if you use the wider Intercom suite, product tours, in-app messaging and lifecycle email, you are buying several tools in one invoice rather than overpaying for chat. Whether that describes you is a separate question, and we worked through it in is Intercom worth it.

The same teams on flat pricing

For contrast, here is the other pricing shape. EzyConn has no per-seat fee and no per-resolution fee. AI runs on GPT-4o and Claude and is included on every plan, including the free one. You pick a plan based on conversation volume, and the price does not move when the AI gets better.

Same teamEzyConn planMonthly total
5 agents, 250 AI conversationsBasic (up to 10 seats)$59
10 agents, 500 AI conversationsBasic (up to 10 seats)$59
25 agents, 1,250 AI conversationsEnterpriseTalk to us

Plans run free, then $19, $29, $59, $89 and $189 a month, and annual billing takes 20 percent off. Up to 10 seats start at $59, so the part-time agent problem stops existing. Full detail sits on the pricing page and the feature-by-feature version is on the Intercom alternative page. If you have already decided, the mechanics are in switching from Intercom.

Common questions about the bill

How much does Intercom cost per month?

There is no single number, because two meters run at once. Intercom Essential is $39 per seat per month on monthly billing, so ten seats is $390 before any AI usage. On top of that, Fin is billed at $0.99 per resolution, which is a conversation the AI closed on its own. A ten-seat team whose AI resolves 500 conversations pays roughly $885 for the month. These figures were checked in July 2026.

What counts as a Fin resolution?

A resolution is a conversation the AI finished without a human taking over. It is not the same as a message or a chat. One long back-and-forth that the AI closes on its own is one resolution and one charge, at $0.99. A conversation the AI hands to a person is a different outcome. This matters for forecasting, because your resolution count tracks how good your help content is, not how many people you employ.

Why does the Intercom bill go up when the AI performs better?

Because the AI is charged per resolution. Improving your help centre makes the AI able to close more conversations by itself, and every one of those closures is another $0.99. The saving lands in your staffing budget while the cost lands on your software invoice, and those are usually two different budgets with two different owners. Teams often notice this the first month after a big documentation push.

How do I estimate my own Intercom bill before committing?

Multiply your seat count by the per-seat price of the plan you want, then add $0.99 for every conversation you expect the AI to close on its own. Do the sum three times: a quiet month, a normal month, and your worst month, because peak season is when a per-resolution meter hurts most. Then double the AI number and look again, since that is what happens when the automation starts working properly.

What would the same team pay on EzyConn?

EzyConn charges one flat monthly price with no per-seat fees and no per-resolution fees. Plans run from a free forever tier to Starter at $25 a month and Professional at $95 a month, with 20% off on annual billing. A ten-agent team resolving 500 conversations a month fits on Professional, and the price does not move if the AI resolves more.

Run your own numbers first

The free plan covers 2 seats and 500 messages a month, forever. That is enough to see real deflection numbers on your own traffic before anyone signs anything.

Start free

Intercom figures quoted here are the published Essential seat price and Fin resolution price as checked in July 2026, on monthly billing. All totals are our own arithmetic on a stated assumption of 50 AI resolutions per agent per month, not quotes from Intercom. Verify current rates before budgeting.