How a US physician group halved the time its billing team spent getting claims out the door, without changing its EHR or billing workflow.
Nalashaa partnered with a mid-sized multi-specialty physician group in the United States whose billing team prepared and submitted claims to a number of commercial and government payers through a mix of clearinghouse batches and payer web portals. Charges came out of the group’s EHR and practice management system.
As claim volume grew, submission turned into a manual relay. Leadership wanted to shorten that cycle without adding staff.
The objective was to take the repetitive preparation and submission work off the billing team and show the saving in minutes per claim.
Submission is where small delays and small errors compound. The 2023 CAQH Index found that providers spend 10 to 12 minutes submitting a claim manually, compared with about five minutes when the transaction is fully electronic.
Every claim passed through several screens before it left the building. Billers pulled ready-to-bill charges from the practice management system, rebuilt the claim for each payer’s format, and keyed it into a portal wherever no batch option existed. A single claim took about 12 minutes of hands-on time, and that figure set the ceiling on how many claims the team could clear in a day.
Each payer has its own required fields, modifier rules and attachment requirements. The group’s most experienced billers knew them, and newer staff learned by watching claims come back. Missing or inaccurate claim data is still the leading driver of denials: 50% of providers named it in Experian Health’s 2025 State of Claims survey, and 41% reported denial rates of 10% or higher.
Medicare fee-for-service claims must be filed within 12 months of the date of service under 42 CFR 424.44, and commercial contracts commonly allow far less, often 90 to 180 days depending on the payer. For Medicare, a submission that is incomplete or invalid is returned to the provider and is not treated as a filed claim, so an error found late can push a claim past its deadline. Backlog and bounce-backs were both direct threats to revenue.
Several payers offered only a web portal. Each had its own login flow, screen layout and confirmation format, and none could be fed directly from the billing system. Any automation had to work through the screen the way a biller does.
Claims carry protected health information, and every submission needs a record of who sent what and when. The group would not accept bots that decided on their own.
The group did not need a new billing platform. It needed the repetitive work between “ready to bill” and “received by the payer” taken off its billers’ desks.
Nalashaa built a pipeline of five stages that takes each claim from ready-to-bill to the payer’s acknowledgment of receipt, with software bots running every stage except the exceptions.
An unattended bot reads ready-to-bill claims from the practice management system work queue and gathers the demographic, insurance, diagnosis and procedure data each payer needs. It then builds an X12 837 file or the portal field set.
Every claim is checked against a payer-specific rules library before anything is sent. The checks fall into four groups.
Clean claims go out through the channel each payer supports. Bots upload 837 batches to the clearinghouse and enter portal claims screen by screen. They then record the payer’s acknowledgment that it received the claim, which is the 277CA for EDI claims (the 999 only confirms the file passed format checks) and the confirmation page for portals, and write the result back to the billing system so claim status is current without anyone checking.
Billers work from one queue that shows only the claims a bot could not send. Each entry names the failed check and the data needed to fix it. Once a biller corrects a claim it re-enters validation and goes out with the rest.
Bots log every action with a timestamp and the claim it touched. Payer portal credentials sit in a vault instead of in scripts, and access is role-based. A run dashboard shows claims received, submitted, held and acknowledged, and email alerts flag a failed run.
The engagement started with a measured baseline. Before any build, Nalashaa timed 500 claims drawn from four weeks of real submissions, so seasonal peaks and the most awkward payers were represented.
Submission time was defined as hands-on staff time from a claim becoming ready until the payer’s acknowledgment that it received the claim was recorded, and that baseline became the reference point for the 50% figure.
The rollout was deliberately staged. The first bot covered the payer with the highest claim volume and the most consistent rules, which gave the team a clean read on time saved. For three weeks that bot ran in shadow mode, assembling and validating claims while billers submitted as usual, and the two sets of results were compared line by line.
Billers then moved to spot-checking bot submissions, and finally to working only the exception queue. Each later payer reused the validation library and added only its own rules and submission path, so it went live faster than the one before.
Portals change without notice, so the portal bots were built to fail safe. If a screen does not match what the bot expects, it stops, holds the claim and raises an alert instead of submitting anything.
Billers stayed close to the tuning loop throughout. Their corrections became new validation rules, which is what moved the bots from technically correct to faster in daily use.
Handoff was part of the plan from the start. Rule definitions, bot configurations and portal runbooks were documented as the work progressed, so the group’s billing operations lead could add payer rules and manage the exception queue without returning to Nalashaa for routine changes.
The first production bot went live 10 weeks after kickoff.
The biggest change was not simply replacing portal typing with bots.
The workflow automated the full repeatable portion of claim submission: gathering claim data, validating it against payer rules, sending it through the appropriate channel, recording the acknowledgment, and routing exceptions back to billers.
That reduced manual effort without forcing the physician group to replace its existing EHR or billing systems.
With the submission pipeline and payer rules library in place, the group can extend the same automation foundation into adjacent revenue-cycle activities, including:
Automate the repetitive. Validate before you submit. Keep your billers on the exceptions.
See where bots can remove repetitive work from your claim submission process while keeping exceptions with the people who know how to resolve them.
Talk to a Healthcare Automation ExpertCookies help us deliver our services. By using our services, you agree to our use of cookies Privacy Policy.
