← Back to blog
Automation

How much does business process automation cost in 2026?

Real cost of an automation project for an SME: Zapier, n8n, custom code. Ranges, worked examples, mistakes to avoid.

ThéoThéo · CTO Lucid-Lab
3 min read

The question comes up on every call: "Concretely, how much does it cost to automate a process?"

Short answer: between €1,500 and €30,000, depending on what we automate and with what. The range is wide because "automating a process" can mean copying 3 lines from Google Sheets to Notion, or orchestrating 12 critical systems with complex business rules and error recovery.

Here is how we actually price the work.

The 3 main cost ranges

1. Simple automation: €1,500 to €5,000

A few steps, tools already in place, structured data, low volume.

Typical examples:

  • Sync new leads from a form to a CRM
  • Send an automatic follow-up email 48h after a quote
  • Create an invoice in QuickBooks as soon as a Stripe payment lands

Typical stack: Zapier or Make (paid plans), sometimes a bit of code in Google Apps Script.

Timeline: 5 to 10 days.

2. Intermediate automation: €5,000 to €15,000

Several systems to connect, conditional logic, semi-structured data (PDFs, emails, free-form forms). Often with a touch of AI to understand text.

Typical examples:

  • Extract info from supplier invoices received by email and post them to accounting
  • Automatically qualify inbound leads (scoring + sales routing)
  • Generate meeting notes from recordings

Typical stack: n8n or Make Pro, GPT-4 or Claude for NLP, dedicated database.

Timeline: 2 to 4 weeks.

3. Heavy automation: €15,000 to €30,000+

Critical systems, high volume, complex integrations (ERP, legacy internal tools), reliability requirements (SLA, monitoring, error recovery).

Typical examples:

  • Rebuild a B2B order management workflow with pricing rules
  • Automate the production of personalised quotes from a 5,000-SKU catalogue
  • Industrialise a tender response process with an AI document base

Typical stack: custom code (TypeScript, Python), dedicated database, cloud infrastructure, monitoring.

Timeline: 4 to 12 weeks.

What really drives the price

Three variables matter, in this order:

  1. The state of the input data. If your data is clean and accessible via API, we save 50% of the time. If it's in scanned PDFs or shared Excel files on SharePoint, expect +30%.
  2. The number of systems to integrate. Each system adds time: auth, error handling, tests. Going from 2 to 4 systems rarely doubles the cost, but multiplies it by 1.5x to 1.8x.
  3. Criticality. An internal automation with no direct customer impact can go down for an hour without drama. An automation that touches client billing needs alerts, error recovery, monitoring. That's easily +40% of budget.

The false economies to avoid

"We'll do it with Zapier, it costs €30/month." True, until you have 12 interdependent zaps that are impossible to debug. From 5, 6 connected automations onward, n8n or some custom code is more economical.

"Our intern can do it." Except when they leave, no one knows how it works. Document. Version. Otherwise you'll pay twice.

"We'll wait until it's critical to automate." By then, it's 3x more expensive because you're automating under pressure with no chance to redesign the process.

How we price things at Lucid-Lab

We look at 4 things during the 30-minute Audit Flash:

  1. Volume: how many times it runs per week
  2. Current manual cost: how many hours × hourly rate
  3. Risk of error: what happens if it fails
  4. Available data: where it is, in what format

Output:

  • A cost range (with assumptions)
  • A timeline
  • An estimated ROI over 12 months

No 14-page proposal. No "it depends". If we can't quantify it in 20 minutes, we haven't understood the problem.

One article a month, nothing else.

Real automation and AI cases in SMEs. Zero spam.

Read next