What I do · 04

Automation Requirements

Automation is not always a good idea. Sometimes the investment pays back in four months. Sometimes in thirty years. The difference only becomes visible after the numbers are worked out. This engagement is about doing that calculation before money is spent - and making a conscious decision about whether to proceed.

What this engagement covers

Most automation projects fail not because the technology is wrong, but because no one asked the right question first: is this worth doing at all?

The answer requires two things. First, an honest measurement of current labour costs on the process in question - how much time it actually takes, how often, and what that time costs. Second, a preliminary estimate of what automation would cost - not a precise figure, but a realistic lower bound based on functional requirements and an initial assessment of technical complexity.

Put these two numbers together and you get a first answer: does the saving justify the investment, and over what timeframe? If the answer is no, the engagement ends here - and the client has saved not just the cost of development, but the cost of writing a full technical specification for something that was never going to make financial sense.

If the answer is yes, or promising, the next step is a detailed technical specification and a more precise cost estimate. That is a separate piece of work. This engagement is the gate before it.

How it works

01
Process inventory
We identify which processes or process segments are candidates for automation - where repetitive manual work is concentrated, where errors occur, where volume is high enough to make automation worthwhile in principle.
02
Labour cost measurement
We measure current time expenditure precisely: how long each step takes, how frequently it occurs, and what it costs at the actual salary of the people doing it. This becomes the baseline against which automation is evaluated.
03
Functional requirements
I document what the automated system would need to do - inputs, outputs, logic, exceptions, integrations. This is not a full technical specification. It is enough to communicate the scope clearly to developers and get a realistic complexity assessment.
04
Preliminary cost estimate
Based on the functional requirements, I consult with developers to get an initial complexity assessment and a lower-bound cost estimate. This figure is approximate by design - its purpose is to enable a go / no-go decision, not to sign a contract.
05
ROI calculation & recommendation
I calculate the expected return: time saved, cost saved, investment required, payback period. The output is a clear recommendation - automate, do not automate, or automate partially - with the numbers that support it.
Typical deliverables
  • Process labour cost baseline
  • Functional requirements document
  • Preliminary complexity and cost assessment
  • ROI calculation with payback period
  • Go / no-go recommendation with rationale
Engagement format
Project-based. Typically 1–2 weeks per process depending on complexity and number of sources involved.
Works well for
  • Repetitive manual processes consuming significant staff time
  • Processes involving multiple data sources or formats
  • Situations where automation has been discussed but never costed
  • Before commissioning a technical specification or development
  • When the business case for automation is unclear
Example · Automation feasibility analysis

Three reporting processes - three different answers

A company prepared regular statistical reports for a government authority. Three separate reporting processes were evaluated for automation potential. All three involved collecting data from multiple sources and compiling a structured report. The question in each case was the same: does automating this process make financial sense? The answers were different in each case - which is precisely the point of this engagement.

Cost comparison - annual saving vs investment

The analysis was based on actual time measurements per source, actual salary costs, and a preliminary automation cost estimate of 250 EUR per data source. Labour cost was calculated at 12.50 EUR per hour based on a 2,000 EUR monthly salary and a standard 160-hour working month. Three processes were evaluated: a monthly report with 12 sources at 20–30 minutes per source; a quarterly report with 6 sources at 5 minutes per source; and a weekly report with 10 sources at up to 60 minutes per source - where data arrived as unstructured text in email, requiring significant manual interpretation.

Diagram
Summary - feasibility verdict by report

Report 1 (monthly, 12 sources): current cost 69 EUR/month, automation investment 3,000 EUR, payback 43 months. Verdict: do not automate. Report 2 (quarterly, 6 sources): current cost 50 EUR/year, automation investment 1,500 EUR, payback 30 years. Verdict: do not automate. Report 3 (weekly, 10 sources, unstructured text): current cost 562 EUR/month, automation investment 2,500 EUR, payback 4.5 months. First-year ROI approximately 170%. Verdict: automate.

Key changes
Report 1 - do not automate: payback period 43 months, investment not justified
Report 2 - do not automate: payback period 30 years, process too infrequent
Report 3 - automate: payback in 4.5 months, first-year ROI ~170%
Client avoided spending 4,500 EUR on automation that would not have paid back
One clear automation project identified with strong financial justification

Ready to get things done?

Let's discuss your project - no commitment, no sales pitch.

Get in touch