Business process analysis: understanding together how you work and where software can help

Business process analysis is a consulting service: I spend some time with you and with the people who do the work, we follow an order or a job together from start to finish, and I note where the work gets stuck. Then I suggest what could be improved, based on what you need: sometimes a small change or a link between the programs you already have is enough, sometimes new software is needed, and only in some cases AI. What comes out stays yours, even if you decide not to go ahead.

Does any of this sound familiar?

• The same information is typed into two or three programs, or from a spreadsheet into the management system.
• Answering a customer about a delivery date or a batch takes longer than it should.
• The office system is there, but on the shop floor or in the vans it is hardly used.
• Orders, delivery notes and invoices are retyped by hand.
• Everyone talks about AI, but it is not clear where to start.
• Some things only one person knows how to do.

How does it work, step by step?

1. Free assessment: you tell me about your request and I reply in writing with a proposal and a quote. Up to here there is nothing to pay.
2. On site: a few conversations with the people who do the work (office, shop floor, technicians), an order or a job followed together from start to finish, the spreadsheets and programs you use every day.
3. The map: the steps of each process, who does them and where time is lost.
4. The solutions: for each process I suggest what to improve, based on what you need. Often rules, fields or a link between systems are enough; AI (reading documents, an assistant on your data, forecasts) I only propose where it really saves time. With priorities and costs.
5. You decide: if it makes sense we go ahead in stages, with a working demo every two weeks; otherwise you keep what came out.

When is AI needed, and when not?

It depends on the process. For each one we look at how it works today, what can be fixed without AI and what AI would add, and we pick what is needed. A few examples from real work:

• Customer orders by email. Today they are retyped into the system. Without AI: one order form and checks on the codes. With AI: the order is read automatically and a person confirms it before it goes into the system.
• Supplier delivery notes and invoices. Today they are retyped by hand. Without AI: e-invoice files are imported where they exist. With AI: automatic reading of delivery notes and PDFs, with confirmation.
• Technicians' work reports. Today paper copied out in the evening. Without AI: a report from the phone with photos and signature. With AI: dictated by voice.
• Delivery dates in production. Today the plan is redone by hand for every order. Without AI: scheduling that respects shifts, downtime and phases. With AI: demand forecasting from order history.
• Questions about data ("how much did we sell to that customer this year?"). Today whoever can extract the data answers. Without AI: fixed reports. With AI: an assistant on company data that answers, with a person confirming.

Often the solution without AI is enough: I will tell you before you spend.

Do you keep the system you have?

Usually yes. The office system and accounting stay where they are; what I build connects to them: it reads and writes with your industry software, with e-invoices, with Excel files and with the APIs of systems that have them. At a waste treatment plant, incoming loads are read from the industry software, which keeps handling the official waste registers. On a battery pack line, the MES reads and writes with the existing system.

Some work done

• A plant wholesaler: about 150 supplier invoices a week retyped by hand into the management system, about 4 minutes each, almost 480 hours a year. The prototype reads Excel, PDFs and photos, checks the arithmetic, proposes item and price, and the clerk confirms only the uncertain lines. It runs on the processor they already have. Estimate: three quarters less time, to be verified on real invoices.
• An engineering studio for solar on apartment buildings: the same system used to be drawn three times, in three programs, and each report took about 2 hours. Now study, 25-year business plan, cost split between owners and the meeting documents come from a single record; AI reads the installer's quote and proposes the fields, and nothing is saved on its own. Being set up.
• A shoe manufacturer: customer orders arrive as PDFs, photos or emails; AI reads them (width, colour, quantity per size) and produces the Excel file ready to load.
• A CNC machine shop: hours per job and per activity from a tablet, machine queue, an alert when logged hours do not match clock-ins, analysis of hours per job and operator. AI was not needed here: I took over the system they had and extended it.
• A sheet-metal manufacturer: production scheduling with "what if" simulation and order forecasting; the customer order is read with AI and pre-fills the production order, which a person confirms. Prototype for a European proposal.
• A line making 7,500 battery packs a year: MES with a digital passport for each pack. On the same line, the welds: defects recognised from photos with vision models, studied on more than 23,000 real photos, and instructions to fix them that adapt to whoever is on shift. The European proposal that came out of it scored 12.3 out of 15 (BattWISE, AI visual quality control).
• A waste treatment plant near Padua: traceability of every package with a QR label, connected to the industry software.
• A wood company: batches, bills of materials and production orders, in development.
• A chain of fashion stores with a head office and affiliated stores: one system for stock, sales and documents, in use from 2012 until the chain was sold.
• Electrical and plumbing installation businesses: jobs, site visits, van stock, voice-dictated work report and invoice from the signed report.
• A verified-news website: AI rates each article on ten criteria and the final score is computed by the code, always the same way. Prototype.
• An industrial company with several sites: a security dashboard with an AI assistant.

Some as a freelancer, others as an employee or as head of the development team: the portfolio states the role in each one.

Where does the data end up?

It stays in the company. The software runs on your server and, if needed, the AI models run on your own computers too, without sending documents to providers outside the European Union. Before adding AI we check what GDPR and the AI Act require for that process.

The next step

If you like, write me a few lines about how you work today. I will reply in writing with a proposal and a quote: this part is free, the on-site consulting is paid. On site across the Veneto region, remotely in the rest of Italy and Europe.

Tell me about your request: free assessment and quote.