Vālenz Health applies cash on the day it deposits, with people in control of every posting.
RemitMatch on KORA turns scanned lockbox remittances into matched payment files and a focused exception queue for the accounts receivable team.
- Cash application
- Same day
- Payments applied on deposit day
- Pages per day
- 2,000
- Average daily volume scoped for the workflow
- Posting approval
- Human-led
- The agent prepares files; the AR team authorizes posting
- Confidence and traceability
- Line-level
- Clean matches separated from items requiring research
Overview
A payment arriving in the bank does not finish the work of collecting it. Finance still has to identify which invoice the money belongs to, reconcile the amount, and apply it correctly. When that evidence arrives as scanned remittance advice, the gap between cash received and cash applied depends on how quickly someone can read and match the documents.
Vālenz Health provides integrated self-insurance solutions for employers, health plans, third-party administrators, Taft-Hartley funds, providers, and members. Its solutions span member experience, payment integrity, provider quality, and plan performance. As the business takes on more clients and complexity, its accounts receivable team needs to absorb the associated payment work without a matching increase in manual processing.
The challenge
The work arrives through a bank lockbox as images: hundreds or sometimes thousands of pages of remittance advice, bundled check by check, with layouts that change from payer to payer. There is no consistent data feed to work from. An analyst reads each page, finds the account identifier and payment amount, and searches the open invoice aging report for a match.
Some remittances have no invoice number. The analyst then falls back to patient name and date of service and searches again. Small-print payment amounts are particularly difficult to read, and an incorrect amount or a forced match can create more work than it saves. The analyst has to reconcile the check as well as the individual lines before posting.
While those scans wait, cash remains unapplied and the aging report does not reflect the money already received. A heavy remittance day adds pressure to the close and pulls people away from the exceptions that need judgment. Reading more pages becomes a staffing problem. Vālenz needed a reliable first pass that could prepare routine matches and show the AR team precisely where a decision remained.
The solution
Vālenz partnered with ACTAVA to build RemitMatch on KORA. The workflow has moved from a completed proof of concept into live operation. It takes scanned remittances through reading, extraction, invoice matching, and reconciliation, then returns files that the AR team can review before posting.
One design rule governs the handoff: no dollar on a remittance should disappear into an unexplained match or omission. A payment line either has a supported invoice match or remains visible as an exception with a reason for review. The agent prepares the cash application; it has no authority to post it.
How RemitMatch works
Swipe to see the full diagram
Step 1: Stage the inputs
The scanned lockbox file arrives alongside the current open invoice aging report. Where remittances omit invoice numbers, the workflow also uses a customer-supplied patient lookup.
Step 2: Read every page
High-resolution optical character recognition converts the scans to text, with attention to the small-print payment amounts that drive the matching and reconciliation work.
Step 3: Extract and flag
The agent classifies the pages and extracts account identifiers and payment fields. It flags poor reads as low confidence when it encounters them, rather than carrying uncertainty forward as a fact.
Step 4: Match the invoice
Each payment line is matched to an open invoice. If the remittance has no invoice number, the agent uses the supplied patient and date-of-service lookup; ambiguous matches remain available for research.
Step 5: Reconcile the check
The agent compares the proposed applied amounts with the check total. Each line receives a status: clean match, match with a variance to confirm, or held for research. Variances are explained or raised for review.
Step 6: Hand off for approval
The workflow produces an import file containing the clean matches, a separate exception list, and a plain-English reviewer summary. The AR team reviews the output and authorizes posting.
The agent reads governed inputs from the finance system of record and creates review files. It does not fabricate amounts, identifiers, or matches. Confidence remains visible line by line, and each decision records the source page, the extracted information, the matched invoice, and the reasoning so finance can follow the chain.
KORA holds the processing stages together across large remittance batches. Extraction and matching are evaluated against known document sets, while each run leaves an audit trail. The human approval boundary keeps a prepared import file separate from a posted transaction.
“Our growth depends on absorbing more clients and complexity without a matching cost curve in finance. With ACTAVA, cash applies on the day it deposits, and my team scales with the book rather than behind it.”
The result
Vālenz now applies cash on the day it deposits. The AR team receives prepared matches and a focused exception queue, replacing the initial page-by-page read, manual keying, and invoice search with a reviewable first pass. People retain the decision about what posts and investigate the lines that need more evidence.
The workflow was scoped around an average of 2,000 scanned remittance pages per day. That scope gives the team a consistent process for a large daily document load: read the pages, match the lines, reconcile the checks, and review the exceptions. The move to same-day cash application brings received cash into the applied balance sooner, while the AR team retains responsibility for posting.
For finance, the change is in where attention goes. Routine matches arrive prepared, unexplained amounts remain visible, and the evidence travels with the recommendation. Vālenz can take on more payment complexity while keeping its AR team focused on decisions and exceptions, with approval and traceability built into the work.
More customer stories
- Johns Hopkins MedicineAcademic medicine / AI innovation
Johns Hopkins is building the AI the market won't sell, and benchmarking every agent before it deploys.
- Of workflows no vendor sells
- 90%
- Cheaper per token, owned against rented
- 20–100x
- Tools tested before deployment
- 200+
- Care management
ThoroughCare enrolls patients 6x faster with automated eligibility intake and is targeting an enrollment-rate lift of up to 20%.
- Program enrollment, down from six months
- 30 days
- More enrollments per care manager
- 5x
- Less token burn from agent optimization
- 11x
- Actuarial / Clinical vendor evaluation
Momentum Actuarial scaled clinical vendor evaluation 100x without adding analysts or compromising depth.
- From first agent build to production
- 30 days
- Report output without more analysts
- 100x
Agents you can govern, prove, and own.
Build agents on a platform that tests them before they go live, keeps every action on the record, and puts the intelligence under your control. Orchestrate with KORA, own the model with CURA, and prove every action with the governance built in.