Coopen / Automation / RPA tools
RPA tools: which one you need — and when you need none.
Most comparisons of robotic process automation tools are written by the vendors selling them. This one is written by the people who have to maintain what gets chosen. Here is how the categories differ, what they cost once you include the parts nobody quotes, and how to pick without locking yourself in.
Process & workflow automation
The tool is the last decision, not the first
Teams usually start by shortlisting RPA tools and end up designing their process around whatever they bought. That is backwards, and it is why so many automation programmes stall after two or three bots: the platform fit the demo, not the business.
The order that works is process first, cost second, tool last. Once you know exactly which steps you are removing and what those steps cost per year, the tooling question mostly answers itself — and quite often the answer is smaller and cheaper than the shortlist suggested.
The other thing vendors do not quote is total cost of ownership. Licence per bot is the visible number. The invisible ones are the orchestration server, the developer time to build each automation, the maintenance when a screen changes, and the internal specialist you now need to keep on staff.
What we automate
The categories, and what each is genuinely good at
Four broad options — most businesses end up using two of them together.
- Enterprise RPA suites — Full governance, bot orchestration and audit trails. Worth it at large scale with a dedicated internal team; heavy and expensive below that.
- Open-source RPA libraries — Browser and desktop automation without licence fees. Powerful, but you supply the engineering and the monitoring.
- Workflow platforms — Tools like n8n or Make connect APIs beautifully but do not drive a legacy desktop app. Great for integration, wrong for screen work.
- Custom automation — Code written for your one process. Cheapest to run, most robust, and fully yours — worth it when the process is core to the business.
- Document AI add-ons — Where inputs are unstructured — PDFs, scans, emails — a model reads first and the automation acts on the result.
- Orchestration layer — Whatever you choose, something has to schedule runs, retry failures and alert a human. This is the piece most projects forget.
How to choose
Five questions that settle the tool choice
Run through these before you look at a single pricing page. In our experience they eliminate most of a shortlist in about twenty minutes.
- Does the target system have an API? — If yes, you probably do not need RPA at all — an integration will be cheaper and far more stable.
- How often does the interface change? — Volatile screens punish screen-scraping bots. Stable, old systems are ideal RPA targets.
- What is the volume? — Low volume rarely justifies a platform licence. High volume justifies proper orchestration.
- Who maintains it in a year? — If the answer is "nobody in-house", pick the option with the least specialist knowledge required.
- What does an hour of this work cost today? — That number sets your budget ceiling. If the tool costs more than the work, stop.
The build-versus-licence math
A per-bot licence is a recurring cost that never ends and grows with your usage. A custom automation is a one-off build cost plus modest hosting and occasional maintenance. Crossing over usually happens faster than people expect: for a single well-defined process running at steady volume, custom is frequently cheaper within the first year or two.
That flips at scale. Once you are running dozens of automations across departments, with compliance requirements and several people building bots, the governance and orchestration you get from a real platform stops being overhead and starts being the point.
We have no licence to sell either way, which is why we can give you the honest version: we recommend the cheapest thing that will still be working in three years.
Results
What a good tool decision buys you
In shelfware
No licence for bots that were never built or quietly stopped running.
No vendor lock-in
Documented automations you can move, change or take in-house whenever you want.
Cost to maintain
The right tool for the job breaks less and needs fewer specialists to fix.
FAQ
RPA tools FAQ
Which RPA tool is best?
There is no universal answer — it depends on whether your target systems have APIs, how much volume you run, and who maintains the automation afterwards. For a single high-value process, custom automation usually wins on cost and reliability. For dozens of automations across departments, a platform with real orchestration and governance earns its licence.
Are there free or open-source RPA tools?
Yes. There are mature open libraries for browser and desktop automation with no licence cost. They trade licence fees for engineering time: you build the orchestration, logging and monitoring that a commercial suite includes. That trade is often worth it, and it is frequently what we build on.
Can we start with a free tool and move to a platform later?
You can, if the automations are documented and the process logic is kept separate from the tool. We build that way deliberately, so a change of tooling is a migration rather than a rewrite.
Do we need an RPA developer on staff?
Not for a handful of automations, provided they are documented and monitored. Once automation becomes central to operations, having someone internal who understands it is genuinely useful — and we hand over with that in mind.
How do we know if a process is worth automating at all?
Multiply the minutes it takes by how often it runs by a loaded hourly rate, then add what its errors cost. If that annual number is not comfortably larger than the build and maintenance cost, automate something else first. Our free audit does this math with you.
More automation
Related services
Ready to cut this cost?
Tell us the repetitive work slowing your team down. We'll show you what it costs — and what to automate first.