How to Automate Google Sheets

Google Sheets quietly became the company database

Google Sheets, which plenty of people still call a Google Spreadsheet, is the free spreadsheet in Google Workspace. In most small companies it quietly became the place where the business is actually run. It holds the customer list, the job schedule, the stock count, the commission working and the numbers someone reads out on Monday morning.

It earned that job by being easy to get into. It opens in a browser, several people can type in it at once, and nobody has to ask an administrator for a licence. The bookkeeper, the office manager and the owner can all be in the same tab at the same time, which is more than can be said for most of the software those three were given.

So the spreadsheet sits downstream of every other tool you own. The booking system, the card terminal, the accounting package and the inbox all feed one sheet, because it is the only place all of it can sit side by side.

Every number in it arrives by hand

Almost none of the time spent on a spreadsheet is spent thinking. It goes on moving numbers into it.

Someone opens the booking system every morning, reads yesterday's total off a dashboard, and types it into a cell. Someone exports a report from the CRM, opens the CSV, deletes three columns and pastes the rest into a tab. Someone checks four supplier portals for prices that changed and updates a column by hand. Someone copies the same figures into a second, tidier sheet for a manager who will never open the first one.

Each takes a few minutes, which is exactly why nobody ever fixes them. They become visible only when the person who does them is on holiday and the sheet goes stale for a fortnight, or when a number gets typed wrong and a decision gets made on it.

Google Sheets

Filling the sheet Yesterday's orders, bookings, balances and ad spend, collected from wherever they live and written into the right row before you open the tab.

Working through the sheet Taking each row and doing something with it somewhere else: a reminder sent, a price checked, a status looked up and written back.

Getting the sheet read Turning the tab nobody opens into a short summary that lands in the place your team already looks every morning.

Formulas only reach what is already open to them

The sheet does more on its own than most people use, and it is worth knowing how far that reaches before paying for anything. One sheet can read a range out of another and stay in step with it. A long tab can be filtered down to a short one that updates itself. A table on a public web page can be pulled in and refreshed. Apps Script will run a job on a timer, so a nightly tidy-up or an emailed snapshot really can be set and forgotten.

Every one of those has the same ceiling. They can only reach what is already open to them: another spreadsheet, a public page, a tool with an API and someone on hand to write against it.

The numbers you actually want are almost never in those places. They are behind a login, on a screen, in a supplier portal or a bank or a booking system or a twenty-year-old piece of trade software, sitting there in plain sight with no export button anywhere near it. That is the wall, and it is the reason someone in your office is still retyping figures at nine in the morning.

A spreadsheet that goes and fetches for itself

Change that, and the question stops being what can I import and becomes what do I want to know.

You buy from six suppliers, and every one shows live prices and stock behind its own login. That can be a column instead: this morning's price against what you paid last time, so the line that moved is obvious before you order.

You ship. Tracking numbers go out and somebody spends an afternoon pasting them into courier sites one at a time. Those statuses can be sitting in the sheet by the time you open it, with anything that stopped moving pushed to the top.

The same trick works on anything you can see in a browser. What your competitors are charging today. Which of your listings slipped off the first page. The takings from a system that only ever shows them on a dashboard. Whether the permit you are waiting on has changed status yet. None of it needs an API, a developer, or an integration somebody else has to build first.

One rule holds across all of it: automate the collecting, keep the deciding. A sheet that fills itself overnight is a gift. A sheet that emails your customers on its own is a risk, and should never send without a person saying yes.

An agent does the retyping

Between the portal where the number lives and the cell it belongs in sits a person. Only the deciding needs one.

WebRun is an AI agent that works a real Chrome browser, signed in as you. It opens the supplier portal, reads the price off the screen, and writes it into the right row of your spreadsheet. Because it uses the browser rather than an API, it reaches anything you can reach: no key to generate, no add-on, no integration that has to exist first.

It runs on your schedule in your own private environment. You can watch a run and stop it mid-way, and anything irreversible drafts and waits for you.

The workflows below are already built. Each one names the apps it touches and the steps it takes, so you can read what it does before you switch it on.