Skip to content

What does workflow automation cost, and what drives the price?

Workflow automation costs come in four parts: the one-off build, subscriptions for the platforms and apps it connects, running costs that grow with volume, and maintenance when connected systems change. The number of steps, systems and exceptions drives the build. Volume drives the rest.

What are you actually paying for?

A quote for “an automation” can mean very different things. Split every budget into four parts and compare each separately.

The four parts of an automation budget
PartWhat it covers
BuildMapping the process, connecting the systems, writing the logic and exception handling, testing and handover. Usually one-off.
Platform and connector subscriptionsPlans for an automation platform, if one is used, and any upgrades to connected apps needed for API access. Monthly or annual.
Running costsHosting, task or operation charges, messaging fees and AI model usage. These rise with volume.
MaintenanceMonitoring, fixing breaks when a connected app changes, and adjusting the workflow as your process changes.

This guide explains what moves each part. We do not publish a price list, because two automations with the same name can differ several times over in effort. The AI agent cost guide covers the extra costs when a workflow includes AI steps.

What drives the build cost?

  • Number of systems. Each connected app needs authentication, field mapping, error handling and testing. Three systems is noticeably more work than two.
  • Branching and exceptions. A straight line from form to spreadsheet is quick. A process with approval rules, different routes for different customer types and retries for failures takes longer to build and test.
  • Data quality. If records are inconsistent, with duplicates, free-text fields where you need structured ones or missing IDs, the build must clean or match them first.
  • API quality. Some systems have clear, well-documented APIs. Others offer limited access, require a higher plan, or have no API at all, which may mean a workaround or a different approach.
  • Testing and rollout. Running the automation in parallel with the manual process, then switching over, is part of the build and worth paying for.

For example, a workflow that copies new web-form enquiries into a spreadsheet and emails a confirmation touches two systems in a straight line. One that reads supplier invoices from a mailbox, matches them to purchase orders in your accounting system, routes mismatches for approval and posts the rest is a different scale of project, even though both might be called “automating admin”.

The single biggest lever is scope. Automating the cases that follow the rules and routing the rest to a person is usually far cheaper than automating everything, and often just as useful.

What do connector and platform subscriptions add?

If the automation runs on a no-code platform such as Zapier, you pay for that platform, usually on plans priced by the number of tasks or operations per month. A workflow with many steps, run many times a day, can move you up a plan quickly. Count steps per run times runs per month before choosing.

Connected apps can add cost too. Some CRMs and business tools only allow API access on certain plans, so connecting them might mean upgrading. Check your account’s plan and permissions for each system before the build starts, not during it.

A custom-coded automation avoids the platform subscription but needs hosting, which for most small workflows is modest. Neither route is always cheaper; it depends on volume and how complex the logic is.

Also check per-user charges. If the automation creates work items that staff handle in a shared inbox or app, each of those staff may need a paid seat. Annual billing is often cheaper than monthly, but commits you before you know the volume, so many businesses start monthly and switch once usage settles.

Weigh the cost against the time it saves.

Estimate

Put your own numbers in.

The starting values are an illustrative example, not a benchmark or a neu figure. Use your own volume, the time each one takes today and the share you expect the system to handle without a person.

Hours back each month23.3
Net value each month€600
Over a year€7,200

Time saved is only part of the case: response times, missed work and errors matter too. The ROI guide covers how to set a baseline, and the cost guide what the build and running costs include.

What does it cost to run each month?

Running costs follow volume. Work them out from your own numbers:

  1. Estimate how many times the workflow runs each month, using last month’s records.
  2. Count the billable steps in each run: platform tasks, messages sent, AI calls.
  3. Multiply, then apply each supplier’s current rate for your plan.
  4. Add hosting, storage and any monitoring service.
  5. Add a margin for growth and retries, because failed steps that rerun can count twice.

Messaging is easy to miss. A workflow that sends WhatsApp confirmations incurs Meta’s per-message charges on top of everything else; see the WhatsApp chatbot cost guide. AI steps, such as reading an invoice or classifying an email, add model usage per run.

Why does maintenance cost anything at all?

Automations break for reasons outside your control. A connected app changes its API, an access token expires, someone renames a spreadsheet column, or a new type of invoice arrives in a format the workflow has not seen. None of these are bugs in the original build, but each one stops the work.

Budget for three things: monitoring that alerts a named person when a run fails, a way to see and retry failed runs, and time to adjust the workflow when your own process changes. Whether that is a support agreement, a block of hours or someone on your team, decide it before launch. An unmonitored automation that silently fails is worse than the manual process it replaced.

When is automation not worth the cost?

Automation pays back through volume and consistency. It is often not worth it when:

  • The task happens a few times a month and takes minutes.
  • The process is still changing every few weeks.
  • Each case needs judgement that cannot be written as rules or reliably checked.
  • The systems involved have no usable API and no export.

In those cases, a checklist, a template or a better-organised spreadsheet may be the right answer for now. When automation does make sense, compare quotes on the same scope and ask each supplier to separate the four parts above. To judge the return, use our guide to measuring ROI, and see data entry automation for a common starting point. Our workflow automation service starts with that scoping conversation.

Common questions

Is a no-code tool cheaper than a custom automation?

At low volume and with simple logic, usually yes: no build fee to speak of and a small monthly plan. As volume, steps and exceptions grow, task-based pricing and maintenance time can overtake the cost of a custom build. Compare the two over a realistic period at your expected volume, including the time someone spends keeping the no-code version working.

Do AI steps make an automation much more expensive?

They add model usage per run and, more importantly, testing effort, because AI output varies and must be checked. For tasks like reading varied invoices or classifying free-text emails, that cost is often worth it. If the input is already structured, such as a web form, plain rules are cheaper and more predictable than an AI step.

Who pays the platform and app subscriptions?

Ideally you do, directly, with accounts in your business’s name. That keeps you in control of the data and the workflow if you change supplier. Some providers bundle subscriptions into a monthly fee; if so, ask which accounts exist, who owns them and what happens to them if the arrangement ends.

How long before an automation pays for itself?

It depends on how much staff time and error correction it removes, against the build and recurring costs. Measure the current process first: how often it runs, how long each run takes and how often mistakes need fixing. With that baseline, you can work out a payback period from a real quote rather than a guess.

Bring us the task.

Tell us what happens today, which tools are involved and where it goes wrong. On a free 30-minute review, we’ll work out whether an agent, an automation or neither is the right fix, and what a sensible first release would be.

Start a project.

Two people on a sofa with laptops, planning a project beside a whiteboard of sticky notes

Book a call or send us a message. We read everything.

tash@neu.ie
Working with teams in Ireland and worldwide.

Get in touch

Free 30-minute AI review.