Business Process Automation: How to Identify the Potential

More than 70% of French SMEs carry out repetitive tasks that could be automated, without having any digital tool to measure them. This figure illustrates a common paradox: teams overwhelmed by low-value tasks, managers convinced that "we need to automate something", but no tool to answer the concrete question — *exactly where*, and *for what real gain*?

Business process automation always begins with an identification and measurement phase. Without structured data on your processes, you risk automating the wrong tasks, underestimating complexity or missing the real productivity opportunities.

Why Automation Without Mapping Fails

Poorly targeted automation is a frequent source of disappointment in SME IT projects. The reasons are usually the same:

• You automate exceptions rather than the nominal case. Without a map of the real process, you model the ideal process — and the robots encounter unforeseen cases in their first week.
• You ignore application dependencies. A process that looks simple may involve 4 different applications with incompatible data formats. Application mapping reveals these dependencies before the automation project starts.
• You do not measure the return. Automating 2 minutes of processing on a task that occurs only once a month makes no economic sense. Automation potential must be weighted by volume.

The right sequence is simple: map first, measure next, automate with full information. To lay these foundations, see our complete business process mapping guide.

Step 1 — Qualify the Execution Mode of Each Step

The first data point to collect for each process step is its execution mode:

• Manual: performed entirely by a human, with no tool support or minimal support (email, spreadsheet). The most common case in SMEs.
• Assisted: the human is guided by a tool (form, application interface) but still performs the data entry and decisions.
• Automated: the step is triggered and executed without human intervention (script, API, automated workflow).

Manual and assisted steps are the automation candidates. Already-automated steps can still be optimised (reduced latency, improved reliability, monitoring).

Step 2 — Measure Volume and Duration

The automation potential of a step does not depend solely on its execution mode. It also depends on:

• Monthly volume: how many times is this step performed per month?
• Average duration: how long does each occurrence take?

These two data points allow you to calculate automatable hours per month:

`

Automatable hours/month = (duration in minutes × monthly volume) ÷ 60

`

Concrete example: a data entry step in an ERP takes 8 minutes and occurs 350 times per month. That is 2,800 minutes, or 46.7 hours/month potentially returned to the team.

Projected over a year, that is 560 hours. At a fully-loaded hourly cost of £40, the potential gain is £22,400 per year for a single step in a single process — a figure that speaks directly to finance leadership.

Step 3 — Calculate the Optimisation Score

The optimisation score (out of 100) is a synthetic indicator that weights several factors for each process:

• The proportion of manual steps relative to automated steps.
• The weighted volume: a high-frequency process has more optimisation value.
• The duration of manual steps: the longer a manual step, the more it contributes to the score.
• The complexity of application dependencies: the more tools manual steps involve, the higher the automation gain potential.

A high score (above 70/100) indicates a process with strong automation potential, justifying in-depth analysis and a dedicated project. A low score (below 30/100) means either that the process is already well automated, or that it has little volume or complexity.

In UrbaHive, this score is calculated automatically from each step's metadata — no manual calculation required. The management dashboard shows processes ranked by score, allowing you to prioritise improvement projects in minutes.

Step 4 — Identify the Right Type of Automation

Not all processes lend themselves to the same automation solutions. Once the potential is identified, you need to choose the right approach:

Integration automation (API/connectors)

Suitable for steps that consist of transferring data from one application to another. Example: copying an order from the CRM to the ERP on each validation. The condition: both applications have an accessible API.

RPA (Robotic Process Automation)

Suitable for steps that manipulate graphical interfaces with no API available. A robot mimics human actions (clicks, keystrokes). Less robust than API integration, but applicable even to legacy applications. To get the best out of it, the process must be well documented — again, mapping is what makes the project feasible.

Automated workflow (no-code/low-code)

Suitable for approval, notification and routing processes. Tools like Power Automate, Make or Zapier allow workflows to be built without development. Relevant for assisted steps that follow simple rules.

Generative AI and document processing

Suitable for steps involving the reading, classification or summarisation of unstructured documents (inbound emails, paper invoices, contracts). These solutions are progressing rapidly and opening up new automation potential in SMEs.

Step 5 — Prioritise with the Effort/Impact Matrix

Before launching an automation project, position each candidate on an effort/impact matrix:

Impact is directly correlated with automatable hours per month. Effort depends on the maturity of the tools involved (available APIs, data quality, process stability). A process with a high optimisation score but applications without APIs will sit in priority 2, not priority 1.

Detect At-Risk Processes Before Automating

A process with high automation potential may hide operational risks that make it a priority for a different reason: resilience, not efficiency.

Three warning signals to check before any automation project:

1. Single point of knowledge: if only one person masters this process, automating without first externalising the knowledge is risky. Start by documenting the process thoroughly.
2. Process without an owner: who will validate the business rules to implement in the automation? Assign an owner before launching the project.
3. Process not reviewed in 12 months: the automation might implement outdated rules. Verify that the documented process reflects the real process.

These risks are detected automatically in UrbaHive and flagged in the dashboard, before you even start an automation project. To learn more about the process mapping methodology that collects this data, see our 6-step tutorial.

From Score to Budget: Building the Business Case

An optimisation score and automatable hours form the basis of an automation business case:

1. Select the 3 to 5 processes with the highest score.
2. Calculate automatable hours/month for each.
3. Estimate the fully-loaded hourly cost of the teams involved.
4. Calculate the gross annual gain: hours × cost × 12.
5. Estimate the automation project cost (tool, integration, maintenance).
6. Calculate the return on investment and payback period.

This business case can be built in under an hour using the data produced by UrbaHive. It is directly usable in a management committee.

Conclusion

Business process automation cannot be decreed — it is built, step by step, from a precise understanding of what is actually happening in the organisation. The optimisation score and automatable hours calculation transform an intuition ("we waste a lot of time on this task") into an objective, comparable measure.

Start by mapping 3 to 5 representative processes. The data collected will be enough to identify your first automation projects and build a solid business case.

[Measure your automation potential with UrbaHive](https://app.urbahive.com/signup) | See the Process feature | Explore connectors

FAQ — Business Process Automation

Where to start when you have never mapped your processes?

Choose a high-volume process with frequent complaints — that is usually where the potential is most visible. Map it in a workshop with the operational teams, qualify each step (mode, duration, volume, tool), and calculate the automatable hours. This first concrete result often creates the buy-in needed to scale the approach.

Is RPA suitable for SMEs?

RPA is technically accessible to SMEs, but its implementation and maintenance cost is often underestimated. It is appropriate when there is no API alternative. Before any RPA project, check whether the applications involved have recently exposed an API — this is often the case with modern SaaS tools.

How do I calculate the fully-loaded hourly cost for a business case?

The fully-loaded hourly cost includes gross salary, employer contributions and a share of overheads. A simple rule of thumb: annual gross salary × 1.45 (contributions) ÷ 1,600 working hours. For a profile at £35,000 gross annual, that gives roughly £32/hour fully loaded.

Does a high optimisation score always mean you should automate?

No. The score identifies the potential, not the obligation. Other factors come into play: process stability (a process under active redesign is a poor candidate), API availability, and business criticality. The score is a prioritisation tool, not a decision.

What is the difference between automation and digitalisation of a process?

Digitalisation means replacing a paper medium or informal communication channel with a digital tool (online form, workflow in a tool). Automation goes further: it removes human intervention from all or part of the steps, via scripts, APIs or RPA. Digitalisation is often a prerequisite for automation.