Gemini now explains why your Sheets formula broke

Spreadsheet formulas are great until they break, leaving you to decipher a cryptic error code or guess whether you should have used COUNTIF or COUNTIFS. Google is finally addressing that headache.

Nine months after bringing Gemini into Google Sheets, Google has upgraded the AI from basic text and chart tasks to handling complex formula troubleshooting. As reported by The Verge on September 25, the sidebar chatbot can now explain why a formula failed and walk you through how to fix it.

A debugger inside your sidebar

Gemini sits in a panel on the right-hand side of Sheets. When you ask it how to manipulate your data, it serves up suggested formulas alongside step-by-step instructions. If several different functions can achieve the same result, the AI explains the trade-offs between each option.

You can insert suggested formulas directly into your cells with a single click. When tasked with counting responses in a messy spreadsheet, Gemini can automatically pick the right function and point it to the relevant column, saving you a couple of web searches.

It also knows its own boundaries. In one test asking to calculate the total travel distance of wedding guests, Gemini admitted it could not natively calculate driving distances within the sheet. Instead, it laid out two workarounds: write a custom Google Apps Script connecting to the Google Maps API, or use the Haversine formula to calculate straight-line distances using coordinates. It even generated the lengthy trigonometric formula and explained that 6371 represented the Earth's radius in kilometers, noting you could swap in 3959 for miles.

Why it matters

We think this is one of the most practical additions to Google Workspace in recent months.

For everyday users and spreadsheet wranglers, it turns the AI sidebar into a built-in debugging tutor. The real barrier in spreadsheets has never been writing simple sums; it is diagnosing why an intricate formula refuses to work. Having the AI inspect your specific cells, explain the logic, and hand you the fix right where you work eliminates the clunky routine of copying errors into a separate browser tab.

That said, it is not total magic. The AI still cannot automate cross-application heavy lifting, such as pulling external geocoding data for you on the fly. You still have to do some legwork if your dataset is missing core pieces like coordinates.

In the broader productivity race against Microsoft 365 Copilot, Google is making a smart, grounded play. Rather than hyping autonomous spreadsheet agents, focusing on instant error explanations targets the single biggest frustration of working with tabular data.