How a value-based care company replaced an onshore data entry team with computer-use agents
A venture-backed healthcare services company processing ~10,000 hours/year of EHR data extraction.
The customer
Our customer is a venture-backed healthcare services company operating in value-based care. Their core workload is high-volume clinical data extraction from leading EHR systems on behalf of payers and providers. They run dozens of concurrent remote desktop sessions into customer-hosted environments under BAA.
The problem: the true cost of going onshore
Until recently, much of this work ran through an offshore operations team. Labor was cheap, management was lean, and the margins held. That changed when the company picked up state and local government contracts with data residency and workforce requirements that forced the entire operation onshore. They stood up a team of three onshore data entry specialists at $15/hour and immediately hit a wall.
"We won contracts that required onshore processing. The moment we started costing out what that team would look like, we realized we needed a fundamentally different approach."
Three full-time data entry specialists at $15/hour is $93,600/year in base wages. Benefits, PTO, facilities, and training push the fully loaded rate to ~$22.50/hour, closer to $140K/year for three people.
That figure excludes the costs that don't appear on the job posting. EHR-specific SOP training takes weeks before each hire is productive. QA review is needed to catch errors on completed records. Attrition is relentless in a role with almost no career path, and every departure restarts the cycle.
Three FTEs could process a fraction of the volume the business was generating. Scaling to meet demand meant hiring more specialists, and onshore data entry hires take roughly two months from posting to full productivity. By the time a new hire was ramped, the backlog had grown again. Growth created operational debt faster than the team could absorb it.
The unit economics that worked offshore were structurally broken onshore.
Each new contract added records, each record added hours, and each hour added headcount, with no efficiency gains over time. The company needed a way to handle 10,000 hours of annual processing volume without doubling its onshore team and without waiting months to hire and train each new specialist.
01. Cost
64% lower
Cost per automation hour
Before
$22.50
After
~$8
02. Headcount
0 ops FTEs
Operations headcount required
Before
3 onshore FTEs
After
0
03. Capacity
1-2 days
To deploy new automation capacity
Before
~2 weeks
After
1-2 days
The solution: Nen computer-use agents
The company replaced its manual operations workflow with Nen's computer-use agents running against the same remote desktop sessions. Three things mattered:
Instant scalability without hiring. Agents handle the same EHR workflows (form navigation, field entry, record extraction) across as many concurrent sessions as the workload requires. Scaling volume means deploying more agents.
No specialist required. Nen automations are Python-driven through an MCP server. Any software engineer already on staff can build, iterate, and debug an automation.
Smaller compliance surface. Every person touching PHI is a BAA scope expansion, an annual training requirement, an access review, and an audit trail. Removing the ops floor removed an entire category of recurring compliance work. The remaining surface is a per-workspace configurable telemetry layer.
| Metric | Before | After |
|---|---|---|
| Cost per hour | $22.50 | ~$8 |
| Ops team | 3 FTEs | 0 |
| Time to add capacity | ~2 weeks (hire and train) | 1-2 days (build automation) |
Total cost of ownership
The cost math behind the headline figure:
| Line item | Before (onshore manual) | After (Nen agents) |
|---|---|---|
| Labor (base wages) | $15/hr | - |
| Fully loaded labor | $22.50/hr (benefits, PTO, facilities, training) | - |
| Nen platform and orchestration | - | Included in modeled blended automation cost |
| Model inference | - | Included in modeled blended automation cost |
| Engineering build | - | ~10 engineering hours per automation |
| Per-hour cost | $22.50 | ~$8 |
Pricing note: Nen's public cloud desktop pricing is separate from managed workflow pricing. The ~$8/hour figure here is a deployment-specific blended automation cost model for this customer, including workflow infrastructure and model usage. Public cloud desktop usage is listed at $1.20/hour.
The modeled managed workflow cost includes the infrastructure around desktop automation: API gateway, queues for serialization, observability, and scheduling. These are features that would take an internal engineer months to build from scratch.
If you're running onshore data entry against EHRs and the math is breaking, reply to this email and we'll walk through your numbers.