Show your work

"Can I trust an agent's number?"

It's the question every analyst eventually asks, usually right before a deliverable goes out. The cell says 78.4%. The agent that produced it is gone — context window flushed, chat archived, prompt forgotten. The model is a snapshot. The reasoning evaporated when the tab closed.

A confident number with no audit trail is worse than no number at all. It leaks trust the moment someone in the meeting asks why.

The diagnostic#

ChatGPT-in-a-spreadsheet is a fluent stranger sitting next to you. It will fill any cell you point at. What it won't do — what no chat-shaped tool can do — is leave behind a structured record of how it got there. A cell comment is a paragraph of prose. You can't search it in a structured way, you can't compare two versions of it, and you can't ask it a question like "which assumptions, if they broke, would move this number?" A note can't do that. A connected web of reasoning can.

The same problem exists without agents. An analyst leaves the team and their model becomes archaeology. The formulas survive; the why doesn't. We've all inherited a workbook whose tabs are named final_v3_USE_THIS and spent a Tuesday reverse-engineering somebody's worldview from cell references.

The fix isn't a better comment system. It's giving reasoning the same first-class treatment we give formulas.

The design move#

A .deepcell carries a structured map of the reasoning right alongside the numbers — the claims, the assumptions, the evidence, and how they connect. Anatomy of a .deepcell covers the full section list; the one we care about here is Reasoning. Four building blocks:

• Claim — a load-bearing statement about the model. Seven kinds: a thesis, a risk, a catalyst, a counter-argument, an open question, the market consensus, or a piece of established knowledge.
• Assumption — an input you're choosing to believe. Each one is marked as holding, uncertain, broken, or superseded.
• Evidence — an anchor to something outside the model: a link to the source (a filing, a web page, a document, or another model), an optional as-of date, and an excerpt.
• Argument — a labeled link between any two of the above: one thing supports, refutes, depends on, derives from, is a variant of, supersedes, contradicts, or references another.

Claims link to the cells they justify. The reasoning map isn't decoration. It's the part of the file that explains the rest of the file.

Here's what a thesis looks like in the wild, slightly fictionalized from a real working model:

Thesis: Gross margin reaches 78% by mid-2027, confirmed by the Q3 call. Assumption it depends on: per-token inference cost falls 30% a year. The link: the thesis depends on that assumption.

The thesis points at two projected quarters of gross margin. It depends on an assumption about inference cost decay. If that assumption breaks, the thesis is the first thing that should be re-examined.

That's the contract: every load-bearing number in the model has a path back to the assumptions that produced it.

What you can do with it#

You can ask the file questions. Ask it: if this assumption breaks, what breaks with it? — and it lists every claim that leans on that assumption, including indirect knock-on effects.

Useful when a quarterly print lands and you want to know, before lunch, which parts of your model the new data point touches.

The more interesting question is the one you ask at the moment you save a version. Flip that inference-cost assumption from holding to broken — say a vendor announces a price floor — and the file walks the reasoning and surfaces every claim downstream of that assumption as a drift candidate. Not automatically falsified. Flagged. The analyst still decides whether the thesis survives the new reality or needs to be superseded by a v3.

Set this to run automatically every time you save a version, and your model can't quietly drift away from its own thesis without someone being alerted.

History is a question too. Ask the file: show me every version of this thesis and what replaced it.

The reasoning survives an analyst leaving. The next person to open the file gets the worldview, not just the worksheet.

Coexistence, not replacement#

An xlsx can carry a cell comment. It can't carry a connected map of the reasoning. That's not a flaw in Excel — comments were designed for the casual reader, the colleague flipping through the tab. They do that job well.

The reasoning map is for the audit trail. The two can live side-by-side: comments stay in the workbook for the casual reader, the reasoning stays in the .deepcell for the auditor, the successor analyst, and the agent running automatic checks. Bring your Excel model in, layer reasoning on top, export back out when you need to.

One aside#

The Reasoning section is the most recent addition to the format — the spec went in last sprint and is the part of .deepcell most actively evolving. The building blocks and relations above are stable; expect more ways to question them to land on top.

Who signs the work#

A connected map of the reasoning is not absolution. The agent can populate claims and evidence at scale, but a claim marked high-strength is still a claim a human is making. The map makes the claim legible — to a reviewer, to a regulator, to the analyst who inherits the model in eighteen months. It doesn't make the claim correct. See the analyst and Claude for how we think about that division of labor: the agent drafts, the analyst signs.

The number on the page is still yours. Now there's a paper trail behind it.

See it for yourself — open a sample .deepcell in the playground. Edit a value, watch the dependents recalculate, inspect the reasoning behind any number.