Customer service software pricing: why we bill by requests, not per seat
Published on September 25, 2026

When you look for customer service software for your shop, you usually compare features first. The price list comes second, and it looks similar at almost every vendor: a price per user per month, three or four tiers, and the feature you actually need sits in the tier above. Only after a few months do you notice that the pricing model has changed your behaviour. You don’t give the seasonal helper their own login because that would be another seat. You don’t switch on automation because it costs extra.
At Humane we chose a different model: you pay by the number of customer requests, in fixed tiers. Every support feature is included in every tier, and there is no annual contract. The returns portal and social media comment replies are separate add-ons with their own price. This article explains why. Not because our model is the best one for everyone, but so you know what to look for when you compare.
The four pricing models you will find
Support software for online shops comes in essentially four pricing models, often combined:
- Per seat. Every person who handles tickets costs a fixed amount per month. Zendesk, for example, bills per agent per month according to its pricing page, Suite Team from €55 and Suite Professional from €115, each billed annually.
- Per ticket. The price follows the number of requests, not the number of people. Gorgias states on its pricing page that its helpdesk scales from 50 to 5,000 tickets a month and is never priced per agent.
- Per resolved conversation. For AI replies a separate line item has become common: an amount per request the AI closes without a human. Both Zendesk and Gorgias list this separately from the base price.
- Per feature. Certain functions, such as automation, reporting or multiple brands, only become available from a higher tier.
None of these models is dishonest. They just create different incentives, and you should know them before you sign.
What per-seat pricing does to a small shop
A price per seat is a clean arrangement for a team with fifteen full-time support staff. For a shop where the owner does support in the morning before packing orders and a helper joins in November, it isn’t.
Run the numbers. Say you get 300 requests a month and answer them yourself. One seat is enough. Then the Christmas season arrives and you bring in two helpers for six weeks. Three seats, for the whole month, even if the helpers only work Tuesdays and Thursdays. In January you cancel the seats again and lose the record of who wrote which reply.
The real problem isn’t the amount. It’s that you start rationing access. The freelancer who handles returns works with your login. The helper gets tickets forwarded by email. Both are bad for security and for oversight. A pricing model shouldn’t push you into sharing logins.
When you pay by requests, the number of people doesn’t matter. With Humane you invite helpers by email, grant permissions individually, such as “approve refunds”, and revoke access when the season is over. That costs nothing extra because it shouldn’t. What that looks like in practice is in Helpers instead of full-time staff.
What per-feature pricing does to your decision
The second pattern that bothers us is the feature staircase. The function that makes the difference sits in the next tier up. With support software that is often the very automation you looked at the tool for in the first place.
The result: you start on the small plan, use the tool as a slightly better inbox, and pay for replies that are still written by hand. Upgrading becomes a project that keeps getting postponed. We wanted the question “do I use automation?” to be an operational decision, not a pricing one. That’s why Humane includes every support feature in every tier: draft mode, automatic reply mode, helper portal, business rules, live parcel tracking. The smallest tier runs the same software as the largest, just with fewer customer requests included.
Why we bill by requests
A ticket is the unit in which support effort actually arises. A shop with 100 requests a month has a different problem from a shop with 2,000, regardless of whether one person or five reply. That’s why the number of requests is the most honest measure for us.
In concrete terms:
- To start: your first 100 customer requests are free. One time, no credit card.
- After that: a flat price per 30 days, tiered by requests. Up to 100 requests €40, up to 150 requests €59, up to 200 requests €79. It goes on in small steps, up to 1,000 requests €339, up to 2,000 requests €549.
- Above 2,000 requests: a custom offer.
All tiers are listed in the pricing overview.
There is no extra charge per additional ticket. If a period goes over your tier, only that period is billed at the higher tier. The next period again counts only what actually arrives. You get a message before you reach the next tier. There is no minimum term, no annual contract and no setup fee. Billing is in euros through Shopify. The returns portal is an add-on you can only book together with Humane Plus, from €19 per 30 days for up to 100 returns.
Only requests that get a reply count. Spam doesn’t. A ticket is a ticket whether the agent answers it automatically, you approve the draft, or you write the whole reply yourself. We don’t charge an additional amount per “resolved conversation”. That is a deliberate difference from the model that has become common for AI replies, and the reason is simple: if every automatically resolved request costs extra, you have an interest in the AI resolving less. We want you to switch it on as far as is right for your shop, not as far as your budget allows.
The math for a typical shop
Say your shop gets 400 requests a month. A third are delivery questions, a quarter are returns, the rest is spread across product questions, address changes and complaints.
With 400 requests you are in the tier up to 400 and pay €153 per 30 days. In November the volume rises to 700 requests. That period is billed at the tier up to 700, so €249. In January you are back at 400 and pay €153 again. You didn’t hire anyone, book a seat or change any settings.
Compare that with a seat model: to handle 700 requests in November you need either more of your own hours or a second person who occupies a seat for the whole month. Billing by requests follows the volume, the seat model follows the staffing plan. For seasonal shops that is a big difference. What a ticket costs when a human answers it is covered in What a support ticket costs.
Where the model has limits
Billing by requests isn’t the best choice for everyone, and we’d rather say so ourselves than have you find out after three months.
- Very many, very short requests. If your support consists of hundreds of one-liners that take ten seconds each, a ticket is a coarse unit. A seat model can be cheaper because one person works through them in series.
- Large teams with stable volume. If you have fifteen full-time support staff, you know your volume and your staff costs. There the seat price is predictable, and the advantage of variable billing shrinks.
- You need to know your volume. If you don’t know how many requests arrive per month, you can’t judge a model based on requests. Count for four weeks first, ideally by category. That is the basis for any support decision anyway.
- Shopify only. Humane is currently available for Shopify only. If you run several shop systems, you won’t find everything you need with us yet.
And there is one more limit that has nothing to do with price: automation has a sensible ceiling. Our aim is to automate up to 90 % of requests, but the remaining cases, complaints, suspected fraud, special requests, need a human. Why we think that is right is in The 90 % limit. A pricing model that rewards you for crossing that line would be the wrong one.
What to check when you compare
Whichever tool you choose, these questions are worth asking before you sign:
- What happens in November? Run your peak volume, not your average, and check whether the price follows the volume or the staffing plan.
- What does automation really cost? If there is an amount per resolved conversation, apply it to your volume and add it to the base price.
- Which feature sits in the next tier? If the answer is “the one I need”, the entry price isn’t your price.
- How long are you tied in? An annual contract for a tool you have never used day to day is a risk. A trial period with real tickets tells you more than any demo.
- What happens with helpers? Ask specifically what an additional login for six weeks costs.
For a broader look at customer service in an online shop, from inbox structure to automation, see the complete guide to customer service for online shops.
In short
Pricing models are not neutral. A price per seat pushes you to ration access. A price per feature pushes you to postpone automation. A price per resolved AI conversation pushes you to hold the AI back. We bill by customer requests because that is the unit in which effort arises, and because it pushes you towards nothing that harms your support. The model has limits, especially for very large teams with stable volume. For most small and mid-sized Shopify shops it fits better than anything else we have seen.
If you want to work out what that means for your volume: try 100 customer requests for free, no credit card, with your real tickets.