What AI automation costs, and how we price it
There's no single sticker price for AI automation, and anyone who quotes one before understanding your workflows is guessing. Here's how we actually price engagements, what moves the number up or down, and how to think about the return, so you can budget with your eyes open.
How our engagements are priced
Most clients move through these stages in order. You can stop after any one of them, and you always know the price before the next step starts.
Free AI assessment
A structured 45-minute session where we map your highest-value automation opportunities and tell you honestly which are worth building. No obligation, no pitch deck.
Discovery & roadmap
A short, fixed-price engagement that scores every opportunity by ROI and feasibility, audits your data and tooling, and hands you a prioritized roadmap with a clear first project and its expected return.
First build
We build and ship one well-defined automation or copilot end to end, typically in two to four weeks, for a fixed, agreed price. You see the number and the expected payback before we start.
Ongoing partnership
Once something works, some clients keep us on to build the next workflows, monitor what's live, and improve it. Entirely optional, month to month, and only while it keeps earning its keep.
What drives the price up or down
Two automation projects can differ ten-fold in cost. These are the factors that actually move it:
Number of workflows
One tightly-scoped automation costs far less than a suite of interconnected ones. We almost always start with a single high-value workflow.
Systems and integrations
Connecting tools with clean, documented APIs is quick. Legacy systems with no API, or ones that need custom connectors, take more work.
How messy the input is
Moving structured data between tools is cheap. Reading unstructured PDFs, scans, or freeform email with an LLM and validating the output takes more building and testing.
Volume and reliability needs
A workflow that runs a few times a day is simpler than one processing thousands of items an hour where every failure has to be caught and retried.
Compliance and governance
Customer-facing or high-risk systems that need EU AI Act classification, documentation, and audit trails carry more work than internal back-office tools.
Human-in-the-loop
The right level of review, approval, and escalation depends on how costly a mistake is. We scope this deliberately rather than guessing.
What's always included, with no hidden costs
- You own everything we build, running in your own accounts and tools.
- Documentation your team can read and change, so you're never locked in.
- We flag ongoing tool and API subscription costs upfront, before you commit, never as a surprise later.
- A live view of what each automation is doing, so the system is never a black box.
The right way to judge the price is against what the manual work costs today. If a workflow quietly burns 80 hours a month across a team, that time has a real cost, and automating it pays back every month, not once.
That's why we lead with ROI, not features. Before any build you get an honest estimate of the hours or errors it removes and the payback period. If the numbers don't work, we'll tell you, and we won't build it.
Pricing questions we hear a lot
How much does AI automation cost?+
It ranges widely with scope: the number of workflows, how many systems you connect, and how messy the inputs are. Rather than quote a number blind, we start with a free assessment and give you a fixed, agreed price for a clearly-scoped first project, with the expected return, before anything is built.
Do you charge a big upfront fee?+
No. We deliberately start small: a free assessment, then a fixed-scope first project that proves value in weeks. You don't commit to a large build until you've seen the first one work.
Is there a monthly retainer?+
Only if you want one. Some clients keep us on month to month to build more and maintain what's live, but it's optional and cancellable. Many engagements are simply a fixed-price build that you then own and run yourself.
Are there hidden costs?+
No. The main ongoing costs are third-party tool and API subscriptions, for example an automation platform or a model provider, and we spell those out before you commit. What we build is yours, with no license fee to keep it running.
What's the cheapest way to start?+
The free assessment. It costs nothing, and often its most valuable outcome is discovering which ideas aren't worth building yet, saving you from spending anything on the wrong project.
Get a real number for your case
The free assessment ends with a clear view of what's worth building and what it would take. No obligation.
Book a free AI assessment