How to spot the lists that should go first
Excel is not the problem. The problem is spreadsheets that have taken on a job they were never built for, in production as much as in the warehouse. Five signs say a list deserves a tool of its own:
- Several people maintain it. It travels by email, and somewhere there is a version called “final_new_2”.
- It is rebuilt from exports again and again. Every week: copy ERP or shop floor data, sort columns, add lookups, format.
- Decisions depend on it. The sequence on the shop floor, a promised delivery date, a purchase order.
- Only one person understands the formulas. When they are ill or on holiday, the list stops.
- Errors show up late. A wrong export or a shifted column only surfaces when someone asks.
The more of these apply, the more the list costs today and the more it pays to replace it.
Which spreadsheets can stay
For many jobs Excel is the right tool and stays that way: a quick calculation, a draft, a one-off analysis, a personal overview. Trying to replace everything wastes time where nothing is lost. And every new tool should offer an export to Excel, because whoever wants to keep calculating should be able to.
Typical candidates in production
- Shift notes and handover list: really a report from confirmations, scrap and downtime. See Automate the shift report.
- Downtime and setup list: really an analysis of machine times per asset. See Analyse machine data.
- Scrap list: really an analysis by line, shift, tool and material. See Analyse scrap.
- Backlog list: really a daily status of open orders by customer and date. See Monitor the order backlog.
- Delivery date sheet in sales support: really an answer on the status of every line item. See Delivery date answers.
- Monthly report from several exports: really a report with fixed definitions and a source for every figure. See Automate the monthly report.
Three routes, and what each is good for
- An ERP module or an MES. Strong when the process follows the standard. Usually a project over months, and the special cases end up in Excel again.
- A BI tool such as Power BI. Good for reports and KPIs. It displays data but rarely takes input, and someone has to build and maintain the data model.
- An application of your own. Fits your process exactly, built by a developer, an agency, with a low-code kit or by AI agents from a description. What matters is who maintains it afterwards.
The routes side by side with public prices: Onagero compared.
How to go about it
- Pick one list and record its effort today: minutes per run, runs per month, people involved. The effort calculator helps.
- Write down where each column comes from: which system, which export, or typed in by hand.
- Get the rules out of the formulas. Colour codes, lookups and filters are rules that so far only live in the file. Write them down in sentences.
- Build the tool that answers the same questions, straight from the source systems instead of exports.
- Run both side by side for a while and compare. Differences often reveal errors in the old list.
- Switch the list off and measure what changed. How is explained in Measuring AI value.
What this looks like with Onagero
The person who maintains the list today describes in their own words what they need. AI agents find the matching data in the ERP, shop floor data or the existing Excel files and build the tool first on a masked copy. It only goes live once a person in your company approves it. By default Onagero only reads from your systems, and Excel stays possible as a source and as an export. More under Connecting Excel lists.