Opinionated workflows by industry: the process comes decided, not as 200 settings
A configurable ERP hands you settings and the job of deciding how you work. An opinionated workflow for your industry comes with statuses and approvals decided.
Read →Guide · How to digitise
Report automation for service firms: how the Monday numbers reach your inbox up to date, without hand-pasted spreadsheets. Four concrete steps, no new tool.
In short
Updated October 2026 · 9 min
Monday, 8:30. The owner of a service firm with 20 people opens three files: quotes sent, projects in progress, unpaid invoices. By lunchtime they have copied columns, phoned a colleague to ask whether the table is current, and arrived at a number they do not fully trust. Report automation does not solve a software problem: it gives the owner back their Monday morning, the time they should spend with clients and thinking about what to sell next. At the service firms we work with, the report is usually the first flow worth automating, because it touches all the others.
A hand-made report costs three things: the time of the most expensive person in the firm, trust in the number they arrive at, and the days lost before it reaches them. The first cost is time. One morning a week is almost 50 mornings a year in which the most expensive person in the firm does copying work. And not just anyone: the one who knows which questions matter.
The second cost is trust. A study that audited 50 spreadsheets in actual use in companies found errors in 94% of them, and in 86% at least one error produced a wrong result (Powell, Baker and Lawson, 2009). When the report is assembled by hand every week, errors do not get fixed: they get copied forward, from "report_final_v2" to "report_final_v3_fixed".
The third cost is delay. The Monday report describes last week. By the time it reaches the owner, the quote that needed a follow-up has gone cold, and the overdue invoice has gained a few more days.
Powell, Baker and Lawson, Journal of Organizational and End User Computing, 2009: 50 operational spreadsheets, audited formula by formula
The reflex of hiring someone to keep the spreadsheets runs into the budget this year, and the starting point of Romanian firms is already low. In the IMM România survey of 24–30 September 2026, across 1,290 entrepreneurs, 35.7% expect to cut staff in the next six months and only 8.5% plan to hire; 53.2% report substantial cost increases (Capital, 2026).
In short: the same volume of work has to be carried by the same people, and margins demand that you know sooner where you are losing money. That is exactly when a late, fragile report costs the most.
And the starting point is low. Eurostat shows that in 2025 only 32% of firms in Romania used specialised management software, meaning ERP (enterprise resource planning), CRM (customer relationship management) or BI (business intelligence), among the lowest shares in the EU, where the average is 53%. Among small firms in the EU, data analysis tools reach only 11%, against 69% among large ones (Eurostat, 2026). For most service firms, the report still lives in Excel, email and WhatsApp.
| Denmark | 73 % |
|---|---|
| Finland | 73 % |
| Spain | 66 % |
| EU average | 53 % |
| Slovakia | 34 % |
| Romania | 32 % |
| Bulgaria | 31 % |
Eurostat, May 2026: enterprises that used ERP, CRM or BI software in 2025
Report automation is not a "business intelligence" project with screens on the wall, but a short chain of four steps, from the questions to the report that arrives by itself.
The report is usually the first flow worth automating, because it touches all the others: quoting, delivery, invoicing. For the full picture of the processes that follow, see the guide to business process automation.
The hand-pasted report
The report that arrives by itself
If you recognise yourself in two of the five signs below, you do not need a new tool; you need the data that already exists to reach one place by itself.
None of the signs says the firm is doing badly. All of them say the same thing: the numbers exist, but in different places, and the only thing linking them is one person, once a week, by hand.
At a service firm with 20 people, the Monday report comes out of the cases the team works in anyway: quotes, projects, invoices, each with a status, an owner and a deadline. In the platform, a request for quote, an audit project or a recruitment mandate is a case with a status and an owner, not a row in a file, and the report is a question asked over the cases, not a document assembled by hand. "How many quotes went out this week and how many were won" is a count on statuses; "which projects are behind" is a comparison between the deadline and today.
Exceptions are rules on the same cases. A quote that has sat in the "sent" status for 10 days without a reply reaches its owner on Tuesday, not the firm's owner on Monday, and the AI Agent drafts the follow-up email from the context of the case. The invoice that goes past due appears on the day it does, with the client and the project it belongs to.
This is what such a firm's report looks like on an ordinary Monday: four answers arrive by themselves, the fifth does not yet have a structured source, and that is exactly where the next step of digitisation begins.
Report automation makes sense when the numbers come from several places and several people; in three situations it changes nothing.
On Monday, before you open the first file, write down on a sheet of paper the questions you want answered and, next to each, where you get the number from today and how long it takes. At the end you have the real map of your report: what is already digital, what is missing and where the morning goes. With this map in hand, the conversation about automation becomes concrete, and the order of the processes that follow the report is in the business process automation guide. If you want to do it with someone who has been through it before, we always start from the firm's real process, not from a tool.
Monday, before you open the first file
Not necessarily. If quotes, projects and invoices each already have a structured place, the report is a link made once between those places, not a new system. A new tool becomes necessary only where the answer does not exist in structured form anywhere: on WhatsApp, on a phone or in someone's head. The map of your questions tells you which case is yours.
It depends on where the answers live, not on the size of the firm. When the sources are already digital, the invoicing software, a project table kept current, a quotes file, the link takes days and the first report arrives the following Monday. When a question has no source, that flow gets digitised first, and the report grows one question at a time.
A dashboard waits for you to open it and assumes you know what to look for. An automated report starts from your questions, arrives by itself at the same hour and, between two reports, sends you only the exceptions. A dashboard is useful once you have the report; before that, it is a nice screen on top of the same hand-pasted spreadsheets.
Give them a place before putting them in the report. If only the coordinator knows how loaded the team is, the report cannot read it; the step is to record the allocation per project in a structured way, with person and week. It is not a detour: every answer that gets a place is an answer you no longer phone about next week.
Once a week for the overall picture, plus the exceptions on the day they appear. A daily report becomes noise nobody reads any more, and a monthly one arrives too late to change anything. Monday at 7:00, with the same questions, and an alert on Tuesday when a quote has gone 10 days without a reply: that is all.
One workflow goes live on a real project, with a success criterion set together.
One email a month, only when we publish. Nothing else.
A configurable ERP hands you settings and the job of deciding how you work. An opinionated workflow for your industry comes with statuses and approvals decided.
Read →AI automates well only what it understands, and it understands only what is clean. Digitisation → context → automation, explained on one real case.
Read →Performance drops in the first weeks after go-live and climbs above the old level from month three. The stages, the critical moment and seven strategies.
Read →