Line-level 835 match, 98% auto-post.
Header-level posting hides the money. Partial payments, takebacks on unrelated lines, and bundling adjustments all look fine when you reconcile at the claim header. They show up at the CPT plus modifier plus rev code level, which is where this engine reconciles. 98 percent auto-post; the 2 percent we cannot auto-post route to the exception queue with the full context attached.
Submitted lines, vs paid lines.
An anonymized sample. A four-line claim for a behavioral health practice. The 837 went out with four service lines. The 835 came back with payment data per line. The engine matches each submitted line to its paid counterpart at CPT plus modifier plus rev code level, reconciles the charge, allowed, paid, and adjustment amounts, and checks contractual variance against the fee schedule.
Line 1 matches cleanly. Line 2 carries a contractual variance outside tolerance. Line 3 was paid but not at the expected rate (partial payment). Line 4 was submitted but missing from the 835 entirely. The engine surfaces all of it. The recovery queue ranks by dollar value.
Header-level reconciliation would have logged this claim as paid $128.40 against $530 charged and moved on. Line-level reconciliation logs $83.60 in recoverable gaps that a specialist can act on: a variance dispute for line 2, a short-pay inquiry for line 3, and a resubmission for line 4. None of those three actions happen if the posting engine stops at the claim header. Multiply across a year of ERAs and the difference between header-level and line-level posting becomes a measurable share of total revenue.
Six capabilities. Line-level honest.
Every 835 ERA runs through the same engine. Line-level matching, contractual variance, partial payment, takeback tracking, bundling detection, payer-specific posting rules, exception queue. No header-level shortcuts that hide the recoverable dollars.
CPT + mod + rev level.
Every paid line on the 835 matches to its submitted line on the 837 at CPT plus modifier plus rev code level. Header-level matching hides partial payments, takebacks on unrelated lines, and bundling adjustments. Line-level matching surfaces all of it.
Short pays, no reason code.
When a line gets paid at less than the expected allowed amount with no adjustment reason code that explains the gap, the engine identifies the short pay and routes the dollar gap to the recovery queue. Distinct from a contractual write-off; this is recoverable revenue.
Allowed vs fee schedule.
For every paid line, compare the payer allowed amount against the expected contractual amount in the fee schedule library. Variance outside tolerance (typically 2 percent) surfaces with expected vs actual delta and payer-specific contract reference. Underpayments accumulate in a dollar-ranked recovery queue.
Cross-period recoupment trail.
Track takebacks across the full payer-provider ledger. A March recoupment against an October claim ties back to the original payment trail. The engine flags high-frequency takeback patterns on specific payer-provider combinations and routes them with the original payment context attached.
NCCI edits, payer rules.
Bundling and unbundling detection by comparing submitted CPT code combinations against paid code combinations. Cross-references NCCI edit pairs and payer-specific bundling rules. Mismatches route to a senior coder for review before posting; compliance and reimbursement issues both caught.
Non-standard codes encoded.
Some payers use non-standard adjustment reason codes. Some require manual contractual write-off entries. Some have specific patient responsibility allocation rules. The payer-specific posting rules library encodes these exceptions and applies them before the auto-post decision. Refreshed when payer policies change.
Four steps. Inbound to posted.
From the moment the 835 ERA arrives from the clearinghouse or direct payer EDI feed, here is the work the engine executes before the dollar amounts land in the patient ledger.
Parse the 835.
X12 835 parser extracts the claim payment information at header and line level. Handles all four standard 835 versions plus the payer-specific extensions. Schema validation flags malformed segments before posting.
Match to the 837.
Line-level match against the originating 837. CPT plus modifier plus rev code. Charge, allowed, paid, adjustment amounts reconciled. Variance check against fee schedule. Bundling check against NCCI and payer rules.
Apply posting rules.
Payer-specific posting rules library applies non-standard adjustment codes, manual write-off entries, and patient responsibility allocation. Auto-post decision: clean cases go to the patient ledger, ambiguous cases route to the exception queue.
Route the exceptions.
Exception queue carries the full context: line-level match output, variance calculation, expected vs actual, recoupment history. Specialist reviews and posts. Recovery queue ranks underpayments by dollar value for follow-up.
Measured outcomes from line-level posting.
Across the active book. Anonymized; individual results depend on payer mix and contract complexity.
Frequently asked questions: 835 reconciliation.
What does line-level 835 to 837 matching actually do?
What is the 98% auto-post rate?
How does contractual variance surfacing work?
What is partial payment detection?
How are takebacks tracked?
What about bundling and unbundling detection?
How do payer-specific posting rules work?
How does this connect to CredPro, the AR workflow, and root cause analytics?
Send 30 days of ERAs. We send back the recovery map.
A free 30-day ERA reconciliation audit. Drop your last 30 days of 835 files and the matching 837 export. We return a four-page audit covering line-level match coverage, contractual variance dollars, partial payment recoverables, takeback patterns, and a 90-day fix plan. A senior partner on the call.