Assemble the history before you start
Watch someone open a stalled claim. Before any decision gets made they find the claim, read what was submitted, find the remittance, look for prior notes, check whether it was resubmitted and when, check eligibility for the date of service, and work out which of several similar claims this actually is.
None of that is judgement. All of it is retrieval, and it is repeated on every touch — including the second and third touch of the same claim by the same person a fortnight apart, because nobody remembers a claim for a fortnight.
Retrieval and judgement behave differently
The distinction is worth taking seriously, because the two halves obey different rules.
Retrieval is deterministic. Given the claim, the steps are the same every time, in the same order, with no decision inside them. It can be done in advance, done for many claims at once, and done by anything able to follow the steps.
Judgement is none of those. It needs the person, it needs their attention, and it cannot start until the retrieval has finished.
That last clause is the important one. Judgement is gated on retrieval, so every minute of retrieval sits inside the expensive half of the process even though nothing expensive is happening. The gating is what makes it worth moving, not the cost of the retrieval itself.
Before, rather than during
Anything deterministic can happen before a person is available. The trigger is not someone becoming free; it is the claim becoming workable — a remittance posts, a hold lifts, a deadline crosses a threshold. At that moment everything the next touch will need can be gathered, whether or not anybody is going to look at it within the hour.
The claim then arrives at a person with its history already assembled, and the first thing they do is decide something.
What counts as assembled is worth defining once rather than per claim. For most stalled claims it is the same short set: what was submitted and when, every response since, every prior touch and its outcome, the eligibility position on the date of service, the deadline that governs, and any sibling claims for the same encounter. Written down, that list is also the specification for what a person should find waiting for them.
This is a change in the order of operations rather than in the amount of work, which is why it is easy to underrate. The retrieval still happens and still costs something. What changes is that it stops occupying the constrained resource — and the constrained resource is attention rather than time.
Two things it also fixes
Consistency. When retrieval happens during the touch, how much of it gets done depends on who picked the claim up and how busy they were. Someone under pressure checks three of the seven things and decides on partial information. Assembled beforehand, the same seven facts are present every time, and outcomes stop depending on who reached the claim.
Interruption. Retrieval is the part most damaged by interruption, because it is a sequence whose state lives in short-term memory. A person interrupted mid-decision returns to the decision. A person interrupted mid-retrieval usually starts again.
Measure the split
Ask a few people to note the moment they stop gathering and start deciding. Not for a month — for a day. The proportion is usually a surprise, and it is the number that tells you whether any of this is worth anything in your operation.
If gathering is a small fraction of the touch, the argument does not apply to you and there are better places to spend the effort. If it is most of the touch, then most of what your team is being paid for is retrieval, which is the work that should not need a person at all.
Our claim-status follow-up page describes what this looks like when the assembly happens before the claim reaches anybody.