Making a RACI matrix that earns its keep
Most RACI matrices are filled in once and never used. The ones that help are the ones where the two structural failures have been checked.
Responsible and Accountable are different people
Responsible does the work. Accountable answers for the result. They are frequently different, and that split is the entire point — if every row has the same name in both columns, the matrix is a task list with extra formatting.
The two failures that matter
Nobody Accountable. Accountability that is not assigned is not held. The row looks complete because someone is doing the work, and then nobody owns the outcome when it slips.
Two people Accountable. Worse than one, and worse in a way that is invisible until something goes wrong: each can reasonably believe the other is carrying it. If you genuinely want joint ownership, say so deliberately rather than arriving at it by not deciding.
These two are worth automatic checking precisely because they are easy to miss by eye on a matrix of thirty rows and eight people.
Consulted has a schedule cost
Consulted means their input is sought before the work is done — a two-way exchange that takes time and can block progress. Informed means they are told afterwards, which costs almost nothing.
Marking people Consulted when Informed would do is one of the more expensive habits in project governance, because every C is a dependency on somebody's availability. A row with six of them is a row that will wait. The useful question for each C is whether their input would actually change what gets done.
When one name dominates the R column
If one person is Responsible for most of the delivery, the matrix has told you something real about risk rather than about formatting. It is a single point of failure, and it is much better discovered while planning than during someone's annual leave.
Your house style may differ, and that is fine
RASCI adds Support. Some organisations allow joint accountability deliberately. Some do not use Consulted at all. A tool that hard-codes one convention and calls everything else an error is wrong more often than it is useful, which is why the rules here are switches rather than assumptions.
What this tool will not do
It calculates from the figures you enter and nothing else. It cannot tell you whether your estimates are honest, whether your dependencies reflect how the work really flows, or whether the plan is achievable with the people you have. Those are judgements, and a schedule that looks precise can still be built on all three being wrong.