Better Questions

I used AI as an editorial assistant to help organize and refine my thoughts; the underlying ideas and personal experiences remain my own.

For years, I maintained what I thought was a pretty good retirement spreadsheet. Like many retirees, it tracked my investments, estimated Social Security, projected taxes and summarized spending. Every year I’d update the numbers, glance through a handful of reports and file it away until the next review.

The spreadsheet faithfully answered every question I asked.

The problem, I eventually realized, was that I wasn’t asking very good questions.

That changed when I began experimenting with Claude for Excel. I wasn’t looking for investment advice or another formula. Instead, I asked it to review my workbook the way an experienced retirement planner might. Which reports were actually useful? Which didn’t help me make better decisions? What information was missing?

The conversation quickly became less about Excel and more about retirement.

One discussion started with what seemed like a worthwhile project: calculating my personal inflation rate. I had years of categorized spending, so it sounded straightforward. But after walking through my data, I realized the calculation wasn’t telling me what I thought it was. My spending hadn’t simply increased. It had changed shape. Commuting disappeared. Travel expanded. Healthcare evolved. I wasn’t measuring inflation. I was measuring the transition from working life to retirement.

That realization sent me in a different direction. Instead of calculating another percentage, I redesigned the report to answer more useful questions. Was insurance quietly becoming my fastest-growing recurring expense? Were healthcare costs tracking my assumptions? Were higher travel expenses really a problem—or simply evidence that I was enjoying retirement? Which large expenditures were one-time events that deserved to be ignored rather than projected into the future?

I found myself making similar changes throughout the workbook. Reports that had once been little more than collections of numbers became decision-support tools. The formulas changed surprisingly little. The questions changed almost completely.

What surprised me most was that Claude rarely “solved” anything. More often, it challenged my assumptions. It asked why I had built a report a certain way or whether a particular metric would ever change a decision. More than once, it convinced me to discard a report I’d spent hours creating because it answered a question I no longer cared about.

Those conversations reminded me of something I’ve learned over the years working with good advisors. Their greatest value isn’t having better calculators. It’s knowing which questions matter.

Most of us already have years of spending records, investment summaries and retirement projections sitting in spreadsheets. We tend to think of them as places to store information. I now think of mine as a framework for thinking. The calculations haven’t changed much. My understanding has.

If you already keep a retirement spreadsheet, try a small experiment. Open Claude for Excel and ask it to review a single worksheet—not the formulas, but the usefulness of the report itself. Ask what an experienced retirement planner would want to see, what decisions the report actually supports and, most importantly, what questions you’re not asking.

You may discover, as I did, that the greatest value isn’t another calculation. It’s finding better questions to ask of the information you already have.