Business process automation: tools and software

In many Italian companies the day opens with the same scene. An alert email from the management software asking for a manual confirmation. A spreadsheet to be copied into another one. A customer asking for the status of an order, with the answer arriving after three handoffs between people.

And yet, according to the ISTAT survey Imprese e ICT 2025, the share of Italian companies with at least 10 employees using artificial intelligence technologies rose from 8.2% in 2024 to 16.4% in 2025 [1][2]. It doubled in twelve months.

This acceleration, however, does not correspond to a uniform result. In Italy, the gap in intensity of use between large companies and small and medium-sized ones widened from about 20 percentage points in 2023 to 25 in 2024 and 37 in 2025 [1]. The technology is there; the effect is missing.

Business process automation is the use of software to carry out repetitive activities in a standardized way — invoicing, order management, internal communications, deadline tracking — reducing human intervention in low-value-added steps.

This article does not offer a product ranking. It explains when automation really pays off, how to prepare for it without "automating chaos," which categories of tools are available to businesses today and how to choose the right tool based on the size of the company. It closes with the most frequent mistakes recorded in the latest OECD reports [4][5].

What automating a business process really means

Is it automation even when a welcome email is sent automatically to a new customer? Technically yes, but calling every single trigger automation is the best way to fail to recognize the real thing when it shows up.

The word "automation" is today among the most used and least defined in business language. Eurostat reports that in 2025, 20% of EU companies with at least 10 employees use AI technologies [3][8], but in many cases these are one-off adoptions — a chatbot, an email trigger — described with different labels. Understanding what automation is and what it isn't is the first condition for not wasting investment on the wrong categories.

The terms most often confused with automation:

A concise operating model: automation = rules + data + execution without human intervention on low-value steps. When one of the three elements is missing — the rules aren't clear, the data is inconsistent, execution requires human judgment — automation produces uncertain or unreliable results.

AI adoption among Italian companies with at least 10 employees has grown significantly: from 8.2% in 2024 to 16.4% in 2025 [1][2]. Among companies that don't use it, 11.5% considered adopting it without following through: in this subset the most cited obstacles are lack of skills (58.6%) and lack of legislative clarity (47.3%) [1]. These numbers suggest that the technology is available before organizations are ready to absorb it.

Recognizing when automation really pays off (and when it doesn't)

How many euros a year does a process cost when it is automated on intuition, without first measuring how many times it runs? Often more than the cost of the software itself, because the mistake only comes to light after the first months of use.

Not all processes benefit from being automated. In the OECD D4SME 2025 survey, process automation is the benefit most expected by small and medium-sized businesses (53% of responses), while the most cited barrier is software maintenance costs (40%), followed by lack of time for training (39%) [4]. The difference between automation that works and automation that turns out to be a hidden cost depends on three variables: repetitiveness, stability of the rules and volume.

The three-variable diagnostic grid:

Repetitiveness — is the process carried out with the same structure every time? High repetitiveness is the prerequisite. If each execution requires a different contextual assessment, automation is not the right tool: you automate the decision, not the decision-making process.

Stability of the rules — are the rules governing the process clear and rarely changed? Ambiguous or constantly evolving rules produce fragile automations that break at the first change. Before automating, it is advisable to stabilize the rules.

Monthly volume of executions — how many times a month is the process carried out? Low volume (fewer than ten executions a month) rarely justifies the investment. High volume (dozens or hundreds of executions) significantly increases the return on investment.

Two contrasting examples:

Recurring invoicing (ideal candidate) — same customer every month, same line items, same VAT logic, same sending channel. High repetitiveness, stable rules, predictable monthly volume. Automation cuts the time from 60 to 5 minutes per cycle.

Handling complex complaints (fragile candidate) — each complaint requires an assessment of the context, empathetic communication and a discretionary decision on compensation. Low repetitiveness of decisions, contextual rules, a strong relational component. Automation can handle the administrative steps (opening a ticket, notification, closing), but not the substance of the response.

The OECD D4SME 2024 report [5] states that, among businesses dissatisfied with their level of digitalization, the most cited cause is lack of time for training (43%), followed by hardware and maintenance costs (37%) and shortage of skills and talent (27%). The first predictor of failure, however, is almost always automating a process that hasn't been mapped yet.

Processes first, then technology: how to prepare for automation

How many automation projects are launched without a written map of the process to be automated? More than you might imagine — and it is the first predictor of failure.

An old rule of common sense says that automating a confused process means getting faster chaos. The twenty-year World Management Survey (Centre for Economic Performance, LSE), based on more than 20,000 interviews in over 35 countries, attributes between a quarter and a third of the differences in total factor productivity across companies and countries to management practices — including formalized procedures and monitoring [6]. The sequence that works has three steps: map the process, simplify it, standardize it. Only then do you choose the tool.

Step 1 — Map the process: before automating, the process must be described in its current form. Business process mapping provides the operational method. Without a map, it isn't possible to identify the steps to automate, the decision points to keep in human hands and the exceptions to handle separately.

Step 2 — Simplify it: mapping almost always reveals redundant steps, unnecessary approvals and duplicated data. Simplifying before automating means not crystallizing inefficiencies in the software.

Step 3 — Standardize it: standard operating procedures (SOPs) turn the simplified process into a documented, repeatable sequence. A written SOP is the functional specification for automation: it tells the system what to do and in what order.

Among Italian companies that evaluated artificial intelligence without adopting it — 11.5% of those that don't use it — the lack of adequate skills is the most cited obstacle, by 58.6% [1]. Part of this obstacle isn't solved by training on the software: it is solved by clarity of the processes upstream. A team that doesn't understand its own manual process won't be able to manage the automated equivalent.

The connection with business systemization is structural: automation is the final phase of the systemization path, not its starting point.

The categories of tools that are useful in a company (what does what, no brands)

Has a company with a good ERP already automated its processes? Having management software is not the same as having automated flows: the ERP manages data, automation manages steps.

Today the software market offers businesses a variety of solutions that, read on a single vendor's website, all look the same. Distinguishing the functional categories clarifies what the company really needs before comparing brand names. The choice changes depending on whether the problem is "reducing handoffs between departments" or "removing copy-and-paste between two systems."

Eurostat (2025) reports that 75.6% of Italian companies use cloud services [3], one of the highest percentages in Europe. The infrastructure is there; the problem is not connectivity, it is the ability to use it to automate operational flows.

The six main functional categories:

1. ERP (Enterprise Resource Planning) — manages the company's operational data: orders, inventory, invoicing, accounting. It doesn't automate flows by itself, but integrates data into a single system that reduces manual duplication. The problem to solve: "my data is scattered across different systems."

2. CRM (Customer Relationship Management) — manages the customer relationship cycle: contacts, opportunities, quotes, communications. It can automate communication sequences (follow-ups, reminders, order confirmations). The problem to solve: "I lose track of business relationships."

3. Workflow and task management tools — organize the flow of work on specific activities: task assignment, progress notifications, approvals. Common tools in the category: Asana, Monday.com, ClickUp, Trello (mentioned only as market examples, not as sources or recommendations). The problem to solve: "I don't know who is doing what and when."

4. Integration platforms (iPaaS) — connect different systems that don't communicate natively, automating data exchange between applications. Common tools in the category: Zapier, Make, n8n. The problem to solve: "my systems don't talk to each other and I have to copy data by hand."

5. Marketing and communication automation — manages communication sequences toward customers and prospects: email marketing, notifications, onboarding communications. The problem to solve: "I can't keep up regular communication with customers without doing it manually every time."

6. Document automation — generates standardized documents (quotes, contracts, reports) from templates and structured data. The problem to solve: "I create every document manually from templates that never get updated."

Choosing the right category depends on the diagnosis of the problem, not on the popularity of the tool. First you identify the specific problem, then you select the category, and finally you compare the solutions within the category.

Choosing the right tool: 4 criteria before price

How many companies choose the tool before writing down who will actually need to use it? The figure isn't available systematically, but in the OECD D4SME surveys lack of time for training is among the most cited barriers, and it is the top one for medium-sized businesses (49%) [4][5].

The price of the software, on its own, is a poor selection criterion. The European Commission's Italy 2025 Digital Decade Country Report notes that only 45.8% of the Italian population has basic digital skills and that the share of ICT specialists in total employment (4% in 2024) remains below the EU average [7], while the OECD D4SME 2025 report ranks lack of time for training second among the barriers to adoption, with 39% of responses [4]. A tool that is perfect on paper but unusable by the internal team is an expense, not an investment.

The four decisive criteria to evaluate before price:

1. Fit with the process already mapped Does the tool solve exactly the problem identified in the mapped process? Or does it require adapting the process to the software's logic? The second situation is common and often produces unsatisfactory results. Test question: "Does this tool handle our real flow, or do we have to change the flow to fit the tool?"

2. Learning curve for the real users Who will use the tool every day? With what level of digital skills? Eurostat reports that in 2024, 72.6% of EU companies with at least 250 employees provided ICT skills training to their staff, compared with 20.8% of those with 10 to 249 employees [9]. This is not a negligible figure: software that requires two weeks of training and a tool you can learn in an afternoon have very different adoption costs. Test question: "Can our team start using it independently within a week?"

3. Integration with existing systems Does the tool communicate with the ERP, the CRM and the other systems already in use? Or does it require an additional layer of manual integration? Every disconnect between systems is a potential point of error and a hidden cost. Test question: "Do we already have a system this tool needs to exchange data with? Is the integration native or does it require development?"

4. Total cost of ownership over three years The monthly price is the starting point, not the relevant figure. The total cost includes: subscription, implementation costs, training, customization, any integrators, and the cost of the team's time for adoption. Test question: "If we add everything up over three years, how much are we really paying?"

Differences by size:

Freelancer (1-3 team members): priority on simplicity and low-code. The tool must work without IT support. The subscription must be negligible compared with the time saved.

Small company (7-15 employees): priority on integration with the main management software. The flow of data between tools is the critical variable.

Mid-sized company (80-100 employees): priority on governance, traceability and scalability. The tool must support audits, access permissions and structured reporting.

The most frequent mistakes in business automation projects

Which mistake costs more: choosing the wrong tool or automating the wrong process? The available data point to the second, because a process error spreads to whatever tool you choose afterward.

The mistakes that derail an automation project repeat themselves with regularity. The OECD D4SME 2024 and 2025 reports identify maintenance costs, lack of time for training and hardware costs as the three recurring barriers for small and medium-sized businesses [4][5]. To these, the growing gap between smaller and large companies recorded by ISTAT in Italy [1] adds a fourth pattern: automating processes that haven't been mapped yet.

The connection with the cluster on continuous improvement is operational: automation projects that fail almost always do so in the post-implementation phase, when there is no correction cycle that makes it possible to identify and solve the problems that emerge in the first weeks.

Limits and conditions of applicability

Business process automation is an effective tool for repetitive processes with clear rules and sufficient volume. It is not a suitable tool for:

Processes with high contextual variability — each execution requires a situational assessment. In these cases, automation can handle the administrative steps but not the decision-making substance.

Rapidly evolving processes — if the rules change frequently, maintaining the automation outweighs the benefit. An updatable checklist is preferable to a rigid system that has to be reconfigured with every change.

Companies without a documentation culture — if processes aren't documented and responsibilities aren't clear, automation amplifies disorder instead of reducing it. The prerequisite is systemization.

The ISTAT data (2025) [1] refer to Italian companies with at least 10 employees, while the OECD D4SME survey (2025) [4] also includes micro-enterprises and self-employed workers. For micro-enterprises with fewer than 5 employees, adoption dynamics are different: operational simplicity and low volume often make basic digitization sufficient, with no need for formal automation.

The European Commission (2025) [7] reports that 70.2% of small and medium-sized Italian companies have reached at least a basic level of digital intensity, a value slightly below the EU-27 average of 72.9% [10]. This figure should be read with caution: an average close to the European one doesn't mean that every company is aligned, nor that speeding up technology adoption is always the right move. The quality of automation — measured by the reduction in errors and operating time — matters more than the speed of adoption.

FAQ

Is it possible to automate without in-house IT skills? Yes, for many categories of tools. Low-code and no-code platforms (integration between applications, document automation, marketing automation) are designed to be configured by people without a technical background. The training required is on the logic of the process, not on programming.

Which process should you automate first? The process with the highest volume of repetitive manual executions and the most stable rules. In many companies, recurring invoicing, automatic sales follow-ups and internal deadline notifications are the most effective starting points.

Does automation reduce jobs? Business process automation in smaller companies mainly concerns low-value-added steps (copying and pasting data, manual notifications, generating standard documents). In almost all the smaller-company contexts observed, the time freed up is redirected toward higher-value activities, not toward headcount reductions.

Operational summary

The automation path in a company follows this sequence: (1) identify the process with the repetitiveness × stability × volume grid; (2) map it with the four-step method; (3) simplify it by eliminating redundant steps; (4) standardize it in an SOP; (5) select the tool category suited to the problem; (6) evaluate the options in the category with the four criteria (process, learning, integration, total cost); (7) implement with a proportionate investment in training; (8) measure the result at 30, 60 and 90 days.

There is one signal that automation is working: the team no longer thinks about the automated step. They carry it out without noticing, because the system handles it reliably.

Conclusion

Business process automation is not a technological shortcut: it is the final phase of a path that starts long before buying any software. The question behind every useful decision is not "which tool to buy," but "which process is clear, repetitive and stable enough to deserve to be automated."

The data collected describe an Italy that is accelerating — AI technology adoption among companies doubled in one year [1][2] — but also a growing gap between those who automate with method and those who do it because it's trendy. The difference doesn't lie in the software chosen, it lies in the process mapping that comes before it.

If you want to start from the right preparation, an operational guide on the subject is available in Business procedures: the operational guide and a complete methodological map in Business process mapping. To frame automation in the broader context of structured growth, the pillar Business management for growing companies is useful.

A company that automates with judgment recovers hours of repetitive execution every week, reduces transcription errors and frees up time for decisions that require judgment. In Italy, the gap with Europe doesn't lie in the basic equipment — 70.2% of small and medium-sized Italian companies reach at least a basic level of digital intensity, compared with 72.9% for the EU-27 average [7][10] — but in the use made of it. That is where, rather than in buying software, the recovery of productivity is decided: in processes that work, not in promises of human resources "freed from bureaucracy."

Automating well is rare because preparing well is rare. And that is exactly where the competitive advantage of the coming years will be decided.

Sources and references

[1] ISTAT, "Rilevazione sulle tecnologie dell'informazione e della comunicazione nelle imprese — Anno 2025", Istituto Nazionale di Statistica, Rome, 2025. Available at: https://www.istat.it/comunicato-stampa/imprese-e-ict-anno-2025/

[2] ISTAT, "Imprese e ICT — Anno 2024", Istituto Nazionale di Statistica, Rome, 2024. Available at: https://www.istat.it/comunicato-stampa/imprese-e-ict-anno-2024/

[3] Eurostat, "Digital economy and society statistics — enterprises", European Commission, Brussels, 2025. Available at: https://ec.europa.eu/eurostat/statistics-explained/index.php?title=Digital_economy_and_society_statistics_-_enterprises

[4] OECD, "SME Digitalisation for Competitiveness — 2025 D4SME Survey, Policy Highlights", OECD Publishing, Paris, 2025. Available at: https://www.oecd.org/content/dam/oecd/en/networks/oecd-digital-for-smes-global-initiative/D4SME-2025-Policy-Highlights.pdf

[5] OECD, "SME Digitalisation to manage shocks and transitions — D4SME 2024 Survey", OECD Publishing, Paris, 2024. Available at: https://www.oecd.org/content/dam/oecd/en/publications/reports/2024/09/sme-digitalisation-to-manage-shocks-and-transitions_735fc44d/eb4ec9ac-en.pdf

[6] Bloom, N., Sadun, R., Van Reenen, J., "World Management Survey at 18: lessons and the way forward", Centre for Economic Performance Discussion Paper, LSE. Available at: https://cep.lse.ac.uk/_NEW/PUBLICATIONS/abstract.asp?index=8859

[7] European Commission, "Italy 2025 Digital Decade Country Report", DG CONNECT, Brussels, 2025. Available at: https://digital-strategy.ec.europa.eu/en/factpages/italy-2025-digital-decade-country-report

[8] Eurostat, "20% of EU enterprises use AI technologies", Statistical news release, December 2025. Available at: https://ec.europa.eu/eurostat/web/products-eurostat-news/w/ddn-20251211-2

[9] Eurostat, "Enterprises that provided training to develop/upgrade ICT skills of their personnel by size class of enterprise", dataset isoc_ske_itts, 2024 data. Available at: https://ec.europa.eu/eurostat/databrowser/view/isoc_ske_itts/default/table

[10] Eurostat, "Digital intensity by size class of enterprise", dataset isoc_e_dii, indicator "Enterprises with at least basic level of digital intensity (DII Version 4)", companies with 10-249 employees, 2024 data. Available at: https://ec.europa.eu/eurostat/databrowser/view/isoc_e_dii/default/table