The HR team that got an afternoon a week back. Each.

The problem

A four-person HR team ran hiring updates the way thousands of teams do: several spreadsheets, none of them quite agreeing, consolidated by hand every week and copy-pasted into bulk emails for hiring managers across the business.

Before any work started, we benchmarked it: 12 hours a week, team-wide. Updates went out late, and copy-paste did what copy-paste does.

This wasn't a guess at what to fix. It came out of an AI review of their operation, ranking the candidates by return. This one sat at the top of the list.

What we built

A Google Apps Script that pulls the team's spreadsheets together, consolidates the information, assembles each hiring manager's update, and drops finished drafts into the team's outboxes.

A person still reads every email before it goes out. The machine does the collating and the typing; the judgment call about hitting send stays human.

The part worth noticing

There is no AI in the running system. Zero tokens, zero per-message costs, nothing to meter.

AI did its work during the build: rapid iteration on Apps Script and Google Cloud tooling turned this into an afternoon of development instead of a week. We used AI as a tool, not as the solution.

Sometimes burning tokens is the wrong thing to do, and an AI consultancy should be the first to tell you so.

The results

From a benchmarked 12 hours a week to under ten minutes: click, review, send. Three hours back per person, every week. Two months in, it runs without babysitting.

The things we didn't put a number on, we'll say in words: fewer copy-paste errors, updates landing on time, and a team of four with their afternoons back.

What didn't go smoothly

The build took an afternoon. Getting access took longer: a few days of back-and-forth to configure cloud permissions and enable the scripting tools. That's the honest shape of corporate automation work, and it's why we scope access before we promise dates. The code is rarely the schedule risk. The permissions are.

The pattern you can borrow

The valuable move wasn't clever technology. It was benchmarking first, ranking the options by return, and having the discipline to ship the boring solution. That review is what we sell as the AI Roadmap, and the build that follows is what we sell as a starter automation.