The most common question I get about AI isn't "what can it do". It's "what will it cost us and when does it pay for itself". And there's a problem with it: while price lists for websites, e-shops or training are a few minutes of searching away, prices for putting AI into business processes are barely published anywhere. The result is that a company gets a quote for €1,600 with no reference point at all for whether that's good value or twice what it should be. This article provides that reference point: three layers of cost, my own prices, three model budgets and a way to calculate payback before you sign anything.
Costs come in three layers: one-off deployment (analysis, building, integration, training), monthly running costs (tool licences + model usage, typically tens to low hundreds of euros a month for a smaller process) and your people's time (checking outputs, mostly in the first weeks). With me, an operations audit costs €1,000 and I deduct it from the implementation that follows, a pilot of one process starts at €2,400 including before-and-after measurement, a program of 3–5 processes starts at €6,000, and optional oversight starts at €320 a month. Payback is trivial arithmetic: hours saved × your real hourly cost. If it doesn't pay back within roughly 12 months, don't do it — or at least not in that shape.
Why you can't find the prices anywhere
There's no conspiracy behind it, just three practical reasons.
Jobs vary wildly in size. "Implementing AI" can mean a week of work on one form or a six-month project wired into a warehouse system. Vendors worry that a published price either scares off small clients or undercuts the large ones.
The market is new and nobody wants to go first. Price lists get compared, and nobody wants to be compared on a product where they're still learning how much work it really takes.
Plenty of quotes rely on vagueness. When a quote doesn't say what exactly will be working at the end, the price can be built on the impression left by a presentation rather than on scope. That's exactly how you pay for a workshop that produces a PDF instead of a working process.
I go the other way: prices are on the service page, and every quote states what specifically will be working when it's done and how we'll know.
Layer 1: the one-off deployment
This is the number you see in the quote. It breaks down roughly like this:
- Analysis and specification (15–25% of the budget). Describing the process precisely enough to automate it: what triggers it, what the steps are, what a good result looks like, who checks it and where the exceptions are. This part gets underestimated most often, and it usually decides whether the thing works.
- Building and integration (40–55%). The automation itself, the prompts and rules, wiring into email, spreadsheets, stock or CRM. The expensive part is usually integrating a system with no decent API — then it has to be done the long way around.
- Testing on real data (15–20%). Not a demo on three sample cases, but a run over dozens of real records, which is where you find out how often the model gets it wrong and on what.
- Training and documentation (10–15%). A tool nobody on the team understands dies within a month. Documentation includes what to do when it breaks.
A quote missing real-data testing and training isn't a cheaper quote. It's an unfinished one.
Layer 2: the running costs everyone forgets
Running costs are a smaller number, but they arrive every month.
- Tool licences. An integration tool like Make or n8n, plus possibly a paid business tier of an AI assistant for the people using it daily. n8n also has a self-hosted variant where you pay for hosting instead of licences — and the data stays with you.
- Model usage. You pay per volume of text processed. For a single inquiry it's fractions of a cent; across thousands of documents a month it becomes a line item that has to be calculated upfront from the real volume.
- Oversight and changes. Processes change: a new type of inquiry shows up, a supplier changes their invoice format, the model gets a new version. Either someone on your side watches it or there's a retainer for it.
For a typical smaller process, licences and model usage stay in the tens to low hundreds of euros a month. But nobody can promise an exact figure without knowing the volume — which is why I calculate it from the numbers that come out of the audit and put it in the quote, not on the invoice.
Layer 3: your people's time
Almost nobody counts this layer, and in the first weeks it's the most expensive one. Somebody has to:
- supply samples of real cases and say what the correct result is,
- check outputs for the first two to four weeks and report what's wrong,
- change their habits — which, for people who've done the job for ten years, is the hardest part.
Budget a few hours a week from one person for the duration of the pilot. If a company isn't willing to give that time, the implementation shouldn't start; it's one of the few reasons I turn work down.
My prices
I keep three steps, each with a fixed price and its own deliverable, so you can stop at any point:
| Step | Price | What you get |
|---|---|---|
| Operations audit | €1,000 | 2 weeks, mapping the agenda with the people who do the work, processes ranked by saving / risk / effort, a time and cost estimate per process, a 6-month plan. Deducted from the implementation if you go ahead within 3 months. |
| Pilot of one process | from €2,400 | 3–6 weeks, baseline measurement, deployment into live operations, integration with your tools, training, documentation, post-deployment measurement and 30 days of tuning. |
| Program of 3–5 processes | from €6,000 | 3–4 months, processes deployed in priority order, internal AI usage rules, a local-model variant when data must not leave the building. |
| Oversight and changes | from €320 / month | Optional. Monitoring, changes when things shift, small extensions. Not a condition for the deployed processes to keep working. |
For companies up to roughly ten people, a full audit usually isn't worth paying for — if the opening call makes it clear where to start, we go straight to the pilot and you save the €1,000. Details and what's included are on the AI automation service page.
Three model budgets
The figures below are scope estimates, not price-list items — the real price always depends on the company, the volume and the state of the systems.
Processing incoming inquiries. An inquiry from email or a form gets sorted, its details extracted and a draft reply and quote prepared. Deployment in the low thousands of euros, running costs low, payback fast — the person who writes every quote from scratch today only checks and sends. The most common first pilot and usually the best benefit-to-risk ratio.
Documents and data entry. Invoices, delivery notes or forms are read and written into the system. More expensive to deploy, because it stands or falls on the integration with your accounting or warehouse software and on scan quality. Running costs grow with volume. Worth it from hundreds of documents a month upward.
Recurring reports and research. A weekly briefing for management, competitor and price monitoring, summaries of long documents. Cheapest to deploy, because it never has to touch operational systems. The saving in hours is smaller but immediate — a good first step for a company that wants to try AI at low risk.
How to calculate payback before you sign
You don't need a twenty-row spreadsheet. Four numbers are enough:
- How many hours a week the activity takes today. Not a desk estimate — log it for a week.
- How much of that actually disappears. Be conservative. Checking outputs stays, so five hours rarely becomes zero — more likely one or two.
- Your real hourly cost for that role, i.e. fully loaded, not net salary.
- Deployment plus running costs for the first year.
Hours saved × 52 weeks × cost = annual benefit. If payback lands within roughly 12 months, it makes sense. If it comes out at three years, either the chosen process is wrong or the quote is expensive. If somebody claims payback within a month, they want a signature, not a result.
What to watch for in a quote
- Percentage savings promised upfront. "You'll save 40% of the time" — before anyone measured what the process costs today — is marketing. The honest version is measuring before and after with the same method.
- Missing running costs. If the quote has no estimate for licences and model usage, you'll learn it from the invoice.
- Accounts in the vendor's name. Access and accounts belong to your company. Otherwise you're also buying a dependency you can't see.
- No documentation or training. Without them you have a black box that nobody can fix once the vendor leaves.
- Vendor commissions. Ask directly. A vendor on commission recommends the tool they get a cut from, not the one that's cheapest to run.
- Deployment without rules. The AI Act has required sufficient AI literacy among people using AI at work since 2 February 2025, and transparency rules for AI-generated content apply from 2 August 2026. Training and written internal rules aren't a nice-to-have.
When I don't recommend it
To be fair to the other side, there are situations where the answer is "not yet":
- The process changes every month and nobody can describe what a correct result looks like.
- The activity takes an hour a month. The automation won't pay for itself in five years.
- The data sits in five tools that don't talk to each other and nobody wants to tidy it up. Then the first step is data cleanup, not AI.
- There's nobody to watch the outputs for the first month.
- It's about decisions on people — candidate screening, employee evaluation or scoring. The AI Act classifies that as high-risk use with separate obligations, and for an ordinary company it isn't worth it.
Summary
- Costs come in three layers: one-off deployment, monthly running costs (licences + model usage) and your people's time for checking. A quote that only mentions the first layer is incomplete.
- My prices: operations audit €1,000 (deducted from implementation), pilot of one process from €2,400, program of 3–5 processes from €6,000, optional oversight from €320 a month.
- Payback = hours saved × 52 × real hourly cost. A sensible target is within 12 months.
- Percentages promised before measurement are marketing. Insist on before-and-after measurement using the same method.
- Accounts, documentation and training belong to you. Without them you're buying vendor dependency.
Frequently asked questions
How much does an AI audit of business processes cost?
With me it's €1,000 for two weeks of work, with a deliverable you can act on even without involving me further: processes ranked by saving/risk/effort, a time and cost estimate for each, and a six-month plan. I deduct the audit fee from the implementation if you go ahead within three months. For companies up to roughly ten people I often recommend skipping the audit and going straight to a pilot.
What are the monthly running costs of AI automation?
They consist of integration-tool licences and model usage billed by volume processed. For a typical smaller process it stays in the tens to low hundreds of euros a month; for agendas with thousands of documents it has to be calculated from real volume. You pay providers directly on your own accounts — it shouldn't pass through a vendor with a markup.
How long until AI implementation pays for itself?
For a well-chosen process, a sensible target is within twelve months. Calculate it as hours saved per week × 52 × the real hourly cost of that role, and compare with deployment plus running costs for the first year. Be conservative about hours saved — checking outputs stays with a human.
Is AI worth it for a company under ten people?
Often more than for a large one — a small team feels every saved hour immediately and has no ten rounds of approvals. The difference is the approach: instead of a full audit, an opening call and a straight pilot of one process is usually enough. Don't start by buying licences for the whole team; start with one process that genuinely slows somebody down.
What's the difference between AI training and AI implementation?
AI training teaches your team to work with AI themselves — prompts, workflows, safety. Implementation is done-for-you: I build the process, wire it in and hand it over working. Most companies need both, which is why training the affected people is part of every deployment. If you want to try AI on a small budget first, start with training.
Do we need a custom application for this?
Usually not. Ordinary operational routine is handled by integration tools and models working over data you already have. A custom application becomes relevant once the process hits a ceiling: proprietary data, complex rules, high volumes or operations the company depends on. I describe the path from the simplest solution to custom code in the guide to your first automations.
Sources and links
- Regulation (EU) 2024/1689 (AI Act) – the AI literacy duty (Art. 4) and transparency rules (Art. 50)
- OpenAI API pricing and Anthropic pricing – per-volume pricing behind model usage costs
- n8n pricing and Make pricing – licence costs for integration tools
- GDPR Regulation (EU) 2016/679 – personal data processing and the processor role