Kill the Step, Keep the Spreadsheet: Where Automation Starts

A founder asked other owners why businesses keep running on spreadsheets, email and messaging apps when better software exists. The most agreed-with answer was short: "Businesses rarely switch to save a few clicks." The cost of switching and retraining dwarfs the benefit. They move off the spreadsheet when it breaks, usually as reconciliation errors or missed invoices that hit revenue.

That answer should change how most automation projects start. The pitch is usually a new system that replaces the mess. The mess, it turns out, is often working.

The spreadsheet is a working system

Another reply described the spreadsheet as "the only thing in the business that bends to how they actually work." Its logic lives in people's heads. The person who knows "if it's this customer we skip the deposit" has never written it down, and the new system has no field for it.

So replacements fail on the exceptions. The demo covers the normal order. The real business is a long tail of special customers, odd suppliers and rules that exist because of one mistake years ago. Staff who have to fight new software to fit the way work happens, every time, stop using it. Someone who used to work at a large corporation described the result. One department refused the new tool, so staff keyed the same data into the old system too, and the two drifted apart.

Find the step people hate

Every business has a step everyone dreads. One owner spent 40 minutes every Friday copying payment figures into a spreadsheet, and hated it. Another described too many Tuesday afternoons cross-checking order numbers between platforms by hand. The reply about the spreadsheet named the pattern: "people switch when you kill a step they hate."

That step is where automation should start. It is specific, everyone agrees it is a waste, and removing it leaves the rest of the workflow alone. Nobody has to learn a new system. The spreadsheet stays, and one painful job stops landing on a person. Adoption follows the relief, and the next step becomes an easier conversation.

This is the practical version of starting boring. The dull project you ship beats the dazzling replacement you argue about.

The steps that slip cost the most

The worst steps are often the ones that quietly stop happening. A founder whose business was growing fast described receivables that normally sat at $5,000 to $8,000. The team was busy with new customers. Nobody had time to chase the overdue accounts. The business was owed more than $25,000 before the owner noticed, while paying the rent.

Follow-ups, chasers and reconciliations are the steps that slip when a team gets busy, and they slip silently. They are also well suited to an agent: the rules are known, the data is in systems you already run, and the result can be checked.

How to pick the first step

Ask the people who do the work which task they would hand off tomorrow. Time it once, so you know what it costs today. Check that the output can be verified: a reconciled total, an invoice sent, a quote filed. Then automate that step and nothing else, and leave the rest of the workflow alone until the team asks for more.

That is how Adam, our operations agent, is deployed. He starts on one mapped workflow, free, takes the repetitive hand-offs such as order entry, renewals and follow-ups, and escalates exceptions with the reason. After the trial he is paid per task completed and verified, so the business pays for the steps that disappear and nothing else.