How to track decisions in a project (so they stop getting lost)
To track decisions in a project, record four things for each one: what was decided, why, when, and who owns it. Keep these in a single findable place (a decision log) instead of scattered across meeting notes and chat, and link each decision to the initiative or task it affects. This turns a vague "we talked about it somewhere" into a record you can search, revisit, and act on.
Why project decisions get lost
Most decisions aren't lost on purpose. They get made in a meeting, a Slack thread, or a doc comment, then move forward without ever being written down as a decision. Weeks later, no one remembers why a choice was made, who agreed to it, or whether it still holds.
The cost shows up as rework and second-guessing. Teams re-debate settled questions, miss the context behind a constraint, or discover too late that the person who owned a decision left or changed their mind. The information exists somewhere, but it's not findable when it matters.
The fix isn't more meetings. It's a lightweight, consistent habit of capturing decisions in a place separate from the discussion that produced them, so the outcome doesn't have to be reconstructed from a wall of notes.
A simple practice for capturing decisions
For every decision worth remembering, capture four fields. What was decided (a clear one-line statement). Why (the reasoning or constraint behind it). When (the date). Who (the owner accountable for it). That's enough structure to make a decision useful months later without turning logging into a chore.
Keep them together in a decision log rather than buried inline. A single, searchable list lets anyone scan past choices, find the one they need, and see the rationale without digging through transcripts. Date-stamping also makes it clear when a decision is stale and due for a revisit.
Then link each decision to the work it affects. A decision attached to its initiative or task carries its context exactly where someone will hit it, so the reasoning travels with the work instead of living in a separate archive no one opens.
How Verkion tracks decisions with the Memory Graph
Verkion is AI project management built on a human-in-the-loop model: the AI proposes, and a person approves before anything is created. Its Memory Graph reads your team's own notes and docs and proposes decisions and action items, each one cited back to the source note it came from, so you can see where a suggested decision originated.
Nothing becomes a tracked record automatically. You review each proposed decision and approve the ones that are real, which keeps the log accurate and free of noise. Suggestions get sharper as the workspace's notes grow, because the graph has more of your team's own material to draw from.
Approved decisions can be linked to your initiatives, workstreams, and tasks, so the rationale lives next to the work it shapes. Your content is processed to generate these suggestions but is not used to train AI models, and each workspace is isolated through Postgres row-level security.
Frequently asked questions
What should a decision record include?
At minimum, four things: what was decided, why (the reasoning or constraint), when it was decided, and who owns it. These fields make a decision understandable months later without having to reconstruct the conversation. Keeping them in a consistent format also makes the log easy to scan and search.
Where should decisions be stored so they stay findable?
Keep decisions in a dedicated, searchable decision log rather than scattered across meeting notes and chat threads. Storing them in one place separates the outcome from the discussion that produced it. Linking each decision to the related initiative or task keeps the context attached to the work it affects.
How is tracking decisions different from taking meeting notes?
Meeting notes capture the full discussion; a decision record captures only the outcome and its essentials, so it stays findable. Notes are where decisions are made, but without a separate log the decision itself is buried in paragraphs. The two work together: notes hold the conversation, the decision log holds the conclusion.
How does Verkion help track decisions?
Verkion's Memory Graph reads your team's own notes and docs and proposes decisions, each cited to its source note. You approve the ones worth keeping (nothing is recorded automatically), and approved decisions can be linked to initiatives and tasks. Suggestions improve as your workspace's notes grow, and your content is never used to train AI models.