I Gave My AI To-Dos One Source of Truth
My AI team and I reconciled 21 personal task records and imported them into Taskwarrior. Checking statuses and priority labels came before the move, followed by a direct view of open work.
Today my AI team and I reviewed 21 personal focus and task records, then brought the reviewed set into Taskwarrior as my canonical personal task manager. The important step was deciding what each record meant before importing it, not simply giving the same list a new home.
At that review, eight items were done, eleven were open, one was cancelled and one was a duplicate. Those distinctions matter: a duplicate is not another commitment, and a cancelled task is not something that should keep returning as unfinished work.
Reconcile before migrating
We imported all 21 reviewed records with their status, importance, effort and notes. Today's work journal records a separate task store and checks that the import did not touch the coaching assistant's existing store.
We also inventoried the places where task inputs and captures already lived. That is a different job from importing records. A successful import establishes what reached the destination; it does not, by itself, resolve every competing list or future capture route.
There was a small but consequential correction to the labels: importance 1 means highest priority. Preserving a number without checking its meaning could faithfully migrate the wrong ordering.
Make the everyday question straightforward
We standardised what happens when I ask to see my to-dos: show all open personal items directly in a concise table. That gives the canonical store a practical conversational use, rather than making it something we maintain but rarely consult.
We also clarified that Taskwarrior's urgency is a weighted sorting score, not priority multiplied by importance. A sorting mechanism and my assessment of importance are not interchangeable labels.
Later, we marked a completed setup task done and refreshed its journal view. That was a concrete update after the initial review, not only a demonstration of importing old records.
Today's result is a reconciled task set, a canonical manager and a clearer way to ask for the remaining work. It is not evidence that I will finish more tasks. That will depend on what happens after the list is correct.
Evidence note: this account follows today's scoped work journal. The recorded import and isolation checks were not independently repeated for this post.