Automated claims intake with fast setup. Here is what fast actually means, and why it is days and not months.

Days, not months. The agent learns by watching a CSR walk through one real claim in the desktop claim client end to end. After that single recording, the recorder generates the structured workflow and the agent runs it. A single intake type is typically in production inside one week, and a multi-state, multi-product intake suite usually ships in two to three weeks. The reason it is days and not months is that there is no per-field selector mapping project. A UiPath ClaimCenter rollout is usually quoted in months because each field has to be mapped, and re-mapped any time the underlying screen changes.

Four steps. One, a CSR opens the desktop claim client and records themselves handling one real FNOL. Two, the Workflow tab in the web app turns the captured screen events into a structured workflow you can read and edit, rather than one you have to write. Three, an ops lead adds validation rules, required-field checks, and the assignment logic; this is where a day or two of the work goes. Four, the agent runs in production, driving the same screens through Windows accessibility APIs. There is no developer-certification course, no per-bot provisioning, and no integration license to buy from the claim-system vendor before you can start.

Because of where the automation reads the screen. UiPath and Automation Anywhere bots are usually built against selectors or pixel matchers tied to the exact layout of the claim form. Building those selectors is the slow part of the rollout, and re-building them every time the form changes is the slow part of the maintenance. Mediar drives the claim client through Windows accessibility APIs, the same interfaces a screen reader uses, so it reads a field as 'edit field, label Claimant Name, value empty' rather than as a pixel region. There is nothing to hand-map at setup, which is what collapses the timeline from months to days.

Yes, and that is the point of using the accessibility layer. Quarterly UI refreshes to Guidewire ClaimCenter or Duck Creek Claims move pixels and re-theme controls, which is what breaks pixel-based and selector-based RPA. The accessible name of a field, the thing a screen reader reads, usually survives a re-theme, so the agent keeps finding the right field. The self-healing layer recovers when a control's accessible name does change, instead of failing the run. The practical effect is that fast setup is not paid back later as fragile maintenance.

The desktop claim client is exactly where it is built to work, and that is the gap most claims-automation tools leave open. Browser-based AI agents work against web-only intake portals, and they are good at that side. The 30 minutes per claim that cost real money is on the other side: opening the Guidewire ClaimCenter or Duck Creek Claims desktop client on a CSR workstation, typing every field, and clicking the assignment rule. That client is a Windows desktop app. Browser-only agents cannot reach it. If your data lives in a desktop claim center, the automation has to live there too.

Mediar bills $0.75 per minute of agent runtime, with no per-seat licensing and no per-bot fee. There is a $10,000 turn-key program fee that converts to credits with a bonus, so it is effectively prepaid usage rather than a separate license. Because capacity is a runtime cost rather than a hiring or licensing cost, a claim-volume spike doubles the minute count you bill, but it does not require a new license, a new workstation image, or a new project from the RPA Center of Excellence.

Mediar is SOC 2 Type II certified and HIPAA compliant, and the desktop agent can run on-prem or inside the carrier's existing Windows VDI rather than in our cloud. PII never leaves the carrier's network unless you choose the managed cloud option. Every run produces an audit log with the field values it typed and a screenshot of the resulting claim screen, which is what compliance and audit teams ask for on each carrier engagement.

The primitive layer is open source. Mediar publishes the Terminator SDK at github.com/mediar-ai/terminator, an open-source library for driving Windows UI Automation from Rust, Python, or Node, and it is the same surface Mediar uses internally. Teams that want a custom workflow or their own agent against the same accessibility-API layer build on Terminator. The Mediar product on top adds the workflow recorder, the no-code web app at app.mediar.ai/web, the SOC 2 and HIPAA control plane, the audit logs, the validation rules, the self-healing layer, and the runtime billing.