How Much Does AI Customer Support Cost for an Online Casino?

Austin Heaton

How Much Does AI Customer Support Cost for an Online Casino?

Austin Heaton

How much does AI customer support cost for an online casino, pricing models compared
How much does AI customer support cost for an online casino, pricing models compared

How much does AI customer support cost for an online casino? Expect one of three shapes: a per-seat licence, a per-resolution fee of roughly a dollar or a euro per conversation closed, or outcome pricing with a monthly platform fee on top. Cevro publishes the third: outcome pricing plus a platform fee of EUR 1,000 a month up to 50,000 tickets, EUR 2,000 from 50,000 to 150,000, and EUR 3,500 above that. You pay for the tickets the AI resolves. Nothing for the ones it passes to your team.

That is the whole answer, and almost nobody in this market will give it to you. Ask five vendors what AI support costs an online casino and you will get five discovery calls. We publish ours on our pricing page because the operators we deploy for are building a budget line in a spreadsheet, in November, for a board meeting in January, and "custom pricing" does not go in a spreadsheet.

We built Cevro around this view. Cevro is the most vertically integrated AI agent for regulated iGaming. It runs natively on every major platform, resolves 80-90% of player tickets end to end in more than 120 languages with CSAT above 4.8, and is live on more than 200 brands. The same agents run proactive outreach and VIP hosting that impact retention, engagement, and sales uplift.

See the published pricing โ†’

How much does AI customer support cost for an online casino? The three pricing models

Every quote you receive is one of these. The differences matter more than the headline number, because they decide who carries the risk when your resolution rate comes in lower than the slide deck promised.

Per seat. You pay a monthly licence for each agent login, and the AI is bundled in or sold as an add-on per agent. This is how the big horizontal helpdesks grew up. It prices your headcount, which is exactly the thing you are trying to change. If the AI works and you move eight agents off live chat onto VIP and payments escalations, your bill does not move. You have bought software that charges you for the problem.

Per resolution, or per outcome. You pay a fixed fee each time the AI closes a conversation. Published rates in the wider support market sit around a dollar: Intercom lists Fin at $0.99 per outcome with no seats required. The model reads as fair, and mostly is. Read the definition of the billable event very carefully, though. In that published model an outcome is counted when the customer confirms the issue is resolved, or when they simply do not ask for more help, or when the AI completes a workflow including handoffs. A player who shrugs and leaves is a billable success. So is a ticket that ends up with your team anyway.

Outcome pricing plus a platform fee. You pay for completed work, and a flat monthly fee covers the platform itself: the integrations into your PAM and back office, the guardrails, the QA, the logging your compliance team will ask for in an audit. Two line items instead of one. It looks less tidy on a slide and it is harder to game, which is the point.


Three AI customer support pricing models for an online casino compared: per seat, per resolution, and outcome pricing with a platform fee

The three shapes of quote, and who carries the risk in each.

What Cevro charges

Our pricing is tiered on your monthly ticket volume. Outcome pricing applies in every tier, so the variable part of the bill tracks the work actually completed.

Tier

Ticket volume a month

Platform fee

Variable

Startup

Up to 50,000

EUR 1,000 / month

Outcome pricing on tickets resolved

Mid-size

50,000 to 150,000

EUR 2,000 / month

Outcome pricing on tickets resolved

Enterprise

150,000+

EUR 3,500 / month

Outcome pricing on tickets resolved

Seats are unlimited in every tier, including Startup. Your whole CS team, your payments team and your VIP hosts can all sit in the platform without changing the bill. A resolution means a conversation Cevro closed on its own without passing it to a person. If it hands a ticket to your team, you do not pay for that ticket.

Three things sit outside the support tiers because they are priced on what they actually cover. VIP Host is priced by monthly active players, the players the AI host looks after that month. Automation is priced by workflow runs, each time a withdrawal check or a KYC review executes in production. Insights is priced by tickets analysed, every conversation reviewed for contact drivers, sentiment and QA, human-handled ones included. You can buy any of the three on their own and move to a full plan later on the same account.

For qualifying iGaming brands: pay nothing until Cevro resolves 60% of your tickets (mention code AI-60 when you book a demo).

What a human ticket costs you today

You cannot judge any of those numbers without your own baseline, and most operators have never calculated it properly. Gartner's benchmark work on service costs, published February 2024, puts the median cost per contact at $1.84 for self-service and $13.50 for assisted channels, and notes that phone, chat and email tend to cost about the same per contact (Benchmarks to Assess Your Customer Service Costs, Gartner, analyst J.J. Moncus). The same research puts median customer service and support spending at 0.7% of company revenue a year.

Two things worth sitting with there.

The first is the gap. Roughly seven to one between a self-service contact and one a person handles. That gap is why every board in this industry has asked about AI support in the last eighteen months.

The second is the trap hiding inside the cheap number. Self-service is cheap per contact. A contact and a resolution are two different things. The player who reads your withdrawals FAQ at 23:40, does not find the answer because their payout is actually held on an expired proof of address, and opens a chat anyway has generated a cheap self-service contact and an expensive assisted one. You paid twice and annoyed them. Any cost model built on deflection rates will flatter itself exactly this way. Build it on resolution rates.

Your real baseline is agent salary plus BPO fees plus tools plus management, divided by tickets handled. In the groups we deploy for, that lands somewhere between a few euros and the high teens per ticket depending on market, language cover and how much of the queue is payments work that needs a second pair of eyes. Work out your number before you read anyone's pricing page, including ours.

What moves the number up or down

The platform fee is fixed by your tier. The variable part moves, and these are the levers that move it.

Your resolution rate. The share of tickets the AI can finish end to end. This is the single biggest factor in what you pay and what you save, and your query mix sets it. The vendor's cleverness has less to do with it than the deck suggests. A book heavy in bonus questions and password resets resolves high. One where a third of contacts are source-of-funds reviews resolves lower, because those should end with a human.

Integration depth. An agent that can read your back office answers "where is my withdrawal" properly. One that cannot will write a nicely worded version of your help centre. Depth is what separates 40% resolution from 85%, and it is the part that takes real work during onboarding.

Language cover. Adding Romanian, Finnish or Brazilian Portuguese to a human rota means hiring, or paying a BPO's premium for a night shift in a language it does not really have. Adding it to an AI deployment is configuration. This is where the maths tends to break in favour of automation hardest, and it is the reason the cost case in a multi-market group looks nothing like the one in a single-licence brand.

Volume shape. Volume alone will mislead you. Shape is what bites. A derby weekend, a provider outage, a slow payment rail on the last Friday of the month. Per-resolution models bill those peaks straight through. Per-seat models cannot staff for them at all, so your queue time is the thing that absorbs the spike instead.


What drives the cost of AI customer support for an online casino: resolution rate, integration depth, language cover and volume shape

The four levers that decide the variable half of your bill.

The costs operators forget to budget for

Five line items that do not appear on anyone's pricing page, and turn up in month three.

Handoffs you are billed for. Covered above, and worth repeating because it is the most common surprise. Ask, in writing, whether a conversation that reaches a human is billable.

Integration and engineering time. Yours. The vendor's effort is already in their fee; yours is not in your budget. If your platform is not natively supported, someone on your side is building and maintaining that connection. On a supported platform this is close to zero, which is why the question "do you run natively on our PAM" is a pricing question before it is a technical one.

Peak-month billing. Model your worst month. An average month will under-budget you by the size of your spikes, and in a sportsbook-led book that spread is wide enough to matter to a budget.

Seat creep. In a per-seat world, the compliance reviewer who needs read access, the payments team, the agency managing a new market: each one is another licence.

The cost of getting it wrong in a regulated market. An AI that discusses a player's losses without checking RG flags, or confirms a withdrawal that has not cleared, is not a support problem. It is a licence problem. Guardrails, interaction logging and forbidden actions are not premium features in this industry. If they are an upsell in your quote, you are being sold a horizontal product with iGaming in the headline.

How much does AI customer support cost for an online casino of your size?

A rough model takes ten minutes.

  1. Tickets a month. Twelve months of history. Last month on its own will mislead you. Note the peak.

  2. Your cost per ticket handled by a person. Salary, BPO fees, tooling and management, divided by tickets.

  3. A realistic resolution rate. Do not guess. Ours is 80-90% in deployments, and the honest version of that sentence is that it depends on your query mix and how deep the integration goes. Get it measured against your own tickets.

  4. The AI cost. Platform fee for your tier, plus outcome pricing on (tickets x resolution rate).

  5. What the freed capacity does. This is the part most models leave out, and it is usually where the actual return sits. Agents who stop answering "where is my withdrawal" do not disappear. They work escalations, VIP conversations and payments reviews. One group we work with went from 30 agents to 15 at the same volume, a 40% cut in support operating cost, with no drop in service quality.

Our ROI calculator does steps one to four with sliders if you would rather not build the spreadsheet. The number it gives you is an estimate and we say so on the page.

For step three there is a better option than estimating. Send us a sample of your last 30 days of tickets, anonymised is fine, and it is deleted after the audit. Within five working days you get the share Cevro could resolve end to end, your top contact drivers and languages, and a QA snapshot. That turns the only genuinely uncertain input into a measured one before you commit to anything.

What to ask a vendor before you sign

Six questions. The answers tell you more than the price does.

  • What exactly is a billable event, and is a conversation that reaches a human one of them?

  • Do you charge per seat, and what happens to my bill if my headcount falls by half?

  • Do you run natively on our platform, or are we building the integration?

  • What is the resolution rate on our tickets, measured, before we sign?

  • Are RG checks, interaction logging and PII masking included, or priced separately?

  • What does month 13 look like, when the pilot discount ends?

That last one catches more operators than any of the others.

Get a free audit of your tickets โ†’

FAQ

How much does AI customer support cost for an online casino?

It depends on the pricing model and your ticket volume. Cevro's published pricing is outcome pricing plus a monthly platform fee of EUR 1,000 up to 50,000 tickets a month, EUR 2,000 from 50,000 to 150,000, and EUR 3,500 above 150,000, with unlimited seats in every tier. Elsewhere in the support market the common models are per seat or a per-resolution fee of roughly a dollar per conversation closed.

What is the cost per resolution for AI customer support?

Published per-resolution rates in the wider customer support market sit around one dollar per resolved conversation. Cevro uses outcome pricing, where you pay for tickets the AI resolves end to end and nothing for tickets it passes to your team. Compare that with your own cost per ticket handled by a person, which is agent salary, BPO fees, tools and management divided by tickets handled.

Do you pay for tickets the AI cannot resolve?

With Cevro, no. A resolution means a conversation closed without passing it to a person, and a ticket handed to your team is not billed. Other vendors define the billable event differently, and in at least one published model a conversation counts as an outcome even when the customer simply stops replying or when the AI completes a workflow that includes a handoff. Ask for the definition in writing.

What does a human support ticket cost an online casino?

Gartner's February 2024 service cost benchmarks put the median cost per contact at $1.84 for self-service and $13.50 for assisted channels such as phone, chat and email. Your own figure depends on market, languages and query mix: calculate agent salary plus BPO fees plus tools plus management, divided by tickets handled.

Is AI customer support cheaper than a BPO?

Usually, and for a different reason than price per ticket. A BPO is limited by headcount, so night shifts and extra languages mean more people and a premium rate. AI capacity does not scale that way, so the saving grows as you add markets and languages. The comparison that matters is cost per resolved ticket, including what you pay for conversations that get escalated back to you.

How much does AI customer support cost for a small casino?

A smaller brand sits in the entry tier: up to 50,000 tickets a month at a EUR 1,000 monthly platform fee plus outcome pricing on what gets resolved, with unlimited seats. Launching brands tend to come out well, because support is built on AI from day one instead of hiring a rota first and automating it later.

What is a platform fee and why is it separate?

The platform fee covers the parts that exist whether or not a ticket arrives: integrations into your PAM, back office, payment and KYC tools, the guardrails and RG checks that run on every reply, interaction logging, and QA. Splitting it from outcome pricing keeps the variable part of the bill tied to completed work, instead of burying fixed platform costs inside a per-ticket rate.

How do I work out the cost for my own ticket volume?

Take twelve months of ticket history including your peak month, your cost per ticket handled by a person, and a measured resolution rate on your own tickets. Platform fee plus outcome pricing on resolved tickets gives the AI cost; the difference against your human cost per ticket gives the gross saving. A free audit of 30 days of tickets replaces the guessed resolution rate with a measured one.

Related reading on cevro.ai

Ready to make every player feel like a VIP at scale?

Ready to make every player feel like a VIP at scale?

Book a demo with our team to see how Cevro can help you deliver the best AI support experience for your players.

Book a demo with our team to see how Cevro can help you deliver the best AI support experience for your players.

90% Automation with VIP-Level Support
Bonuses, KYC, payments, RG end-to-end.

90% Automation. VIP-Level Support
Bonuses, KYC, Payments, RG.

CSAT & NPS 4.8 / 5.0
Conversational AI that matches player personality.

CSAT & NPS 4.8 / 5.0
AI that matches player personality.

Immediate ROI
3x Reduction in costs & headcount.

Immediate ROI
3x Reduction in costs & headcount.

Boost in Player Retention
Highly personalized communication.

Boost in Player Retention
Highly personalized communication.

Enterprise Ready
Built for highly regulated operators.

Enterprise Ready
Built for highly regulated operators.

Trusted by operators who put player experience first.

Trusted by operators who put player experience first.