If your team's decisions live in a dozen Slack threads, you already know the real cost: someone asks "wait, why did we pick this?" three months later, and nobody can find the thread. A decision log fixes that by giving every choice one place to live: the question, the options you weighed, who owned the call, when it was due, and what you actually chose and why.
Here's how to build one in Probe.
Key takeaways
What a decision log tracks
A useful decision record needs more than a title. At minimum it needs: the question being decided, the options considered, who owns the decision, a due date, a status, and the final choice with a short reason for it. Leave any of those out and you'll end up back in chat history trying to reconstruct what happened. The database built below has these as actual fields, not just a notes column.
Step 1: Describe the log with Build with AI
Open Probe, click New Paper, and choose a Doc. Type "Team decisions" as the title. Under the title you'll see Build with AI; click it, and a box opens asking "What is it for?"
Type your request and press Enter. The one used here was:
"A log of our team's decisions. Each decision has the question, the options, who owns it, the date it's due, a status (open, decided, revisit) and the choice we made with a one-line reason. I want a board by status called Open decisions, a table of every decision called All decisions, and a board by owner called By owner."

The Paper Agent takes it from there.
Step 2: Review and apply the proposal
Probe shows a Review required card describing exactly what it's about to build: the fields (Question as the title field, Options, Owner, Due date, Status with Open/Revisit/Decided, Choice made, Reason), the three views, and a row type named "Decision" set as the default. Nothing is created yet.

Click Apply. The database is built for 40 credits (three charges of 13, 14 and 13). Once applied, the database exists with the structure you asked for and no rows in it yet.

Adding an owner to each decision
Owner is a person field, which is what makes the By owner table work later. Every decision needs exactly one owner: the person accountable for making the call and documenting it, not just whoever raised the question. Mapping stakeholders and assigning clear decision rights speeds up decisions, as monday.com (opens another site) notes.
Setting a due date
The Due date field matters for the same reason a deadline matters anywhere. Decision-making works better with a time limit and a clear owner, as sessionlab.com (opens another site) points out. Setting a due date when you log the question, not after, is what keeps "open" decisions from sitting open for months.
Step 3: View every decision in a table
Open All decisions. This is the table view, showing every decision and every field side by side: question, options, owner, due date, status, choice made, and reason. It's the view to scan when you want the full picture at once, or to search for a specific past decision.

Step 4: View decisions by owner
Open By owner. This is a table grouped by owner, so each person's decisions sit together. It's useful in a standup: instead of asking "what are you working on," you can ask "what do you still need to decide."

Logging the choice and why
The two fields that make this log worth keeping are Choice made and Reason. When a decision is settled, fill in what you actually chose and a one-line reason for it. That one line is the thing chat history never preserves well: by the time someone asks "why did we go with this," the reasoning is usually buried under fifty unrelated messages. A one-line reason, attached directly to the decision record, is what you search for later.
Step 2 (continued): Add example decisions
To see the log in action, the next request to the agent was: "Add five example decisions for a small product team, due between November 18 and November 30, 2026, two of them decided with a reason. Start each with 'Example:'."
The proposal added five rows:
The agent noted that Owner is a person field, so it wrote the example names as plain text rather than linking actual accounts, worth checking once applied. Click Apply. This step cost 40 credits (9, 16 and 15), bringing the whole run, database plus example rows, to 80 credits total.

Step 5: Open decisions by status
Open Open decisions, the board grouped by status. The example rows carry Open, Revisit and Decided statuses, and this board groups them into columns by that status. To move a decision from Open to Decided once it's settled, drag its card into the Decided column.

What you end up with
The whole thing, structure plus five example rows, was built across two requests for 80 credits total. From here you delete the examples and start logging your own decisions as they come up, or keep them as a reference for how to fill in the fields.
Questions people ask
What fields should a decision log have?
At minimum: the question, the options considered, an owner, a due date, a status, and the final choice with a one-line reason. These fields, along with Choice made and Reason as separate fields, are what the database built above uses.
Can I track decisions by status and by owner at the same time?
Yes. Because a database in Probe holds one set of records shown through several views, the same decisions appear in the Open decisions board (grouped by status), the All decisions table (every field), and the By owner table (grouped by owner), and an edit in one shows in all of them.
How do I mark a decision as decided?
Drag its card from the Open column to the Decided column on the Open decisions board, then fill in the Choice made and Reason fields on that record.
Does Probe sync this log with Slack or Jira?
No. Probe has no public Paper API and no sync that lets outside tools write into a Paper, so the decision log lives inside Probe rather than updating automatically from chat or ticket tools.
What's the difference between Open, Decided and Revisit?
Open means the decision hasn't been made yet. Decided means it's settled, with a choice and reason recorded. Revisit is the third status option you specified.
