We think this is exactly the kind of manual grind that should have been automated years ago, and it's frustrating that it hasn't been. The user is doing HR work that matters, tracking employee absence patterns, but the process is stuck in a world where someone has to manually count rows, square numbers, and add up days by hand. That is not a skill problem. That is a tool problem. And it is one that a modern spreadsheet, especially one that understands data relationships, can solve in seconds.
The core challenge here is structural. The report exports one row per absence instance, but the Bradford formula requires grouping by person: count the instances per employee, square that number, then multiply by the total days across all their absences. In a traditional spreadsheet, that means writing a formula that can look across multiple rows, identify which rows belong to the same person, and perform a calculation that combines both a count and a sum. That is not a basic VLOOKUP. It is a multi-condition aggregation. And for someone who says they have a basic understanding of Excel, it is genuinely out of reach without help. The fact that an entire HR department has been doing this by hand suggests the organization has not questioned whether the tool is right for the task.
This is where an AI-native approach changes the equation. Instead of forcing the user to learn SUMPRODUCT or array formulas, a spreadsheet that understands natural language can take a request like "for each employee, square the number of absence instances and multiply by the total days" and generate the correct grouped calculation. The user does not need to become a formula expert. They need a tool that understands the logic of their work. The Bradford factor is a straightforward metric, but the way the data is structured makes it needlessly complex to compute. That complexity should be handled by the software, not the person. We believe the real innovation here is not a faster way to type formulas, it is a spreadsheet that can infer the grouping and calculation from a plain description of the problem. That is the difference between managing data and being managed by it.