> For the complete documentation index, see [llms.txt](https://fieldworker.gitbook.io/fieldworker-docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://fieldworker.gitbook.io/fieldworker-docs/fundamentals/pages/ingestion-and-onboarding-agent.md). # Ingestion & Onboarding Agent The Shift to Intelligent Care ## The Automatic Ingestion (AI) Agent The release of our NJ DDD Autonomous Ingestion Agent marks a fundamental shift in how FieldWorker and AbilityHub support care providers. We are moving away from passive "systems of record" where human operators do all the compliance-heavy lifting, and moving into active "systems of intelligence" where the software agents protect the agency's revenue cycle from day one. The Ingestion Agent bridges the gap between static, unstructured documents and structured, actionable database records. It relies on a few core concepts: * **Intelligent Data Ingestion**: The system parses complex, multi-page PDFs (such as the New Jersey Individualized Service Plan) and uses AI to map the text to structured database fields corresponding to the platform's architecture. * **Human-in-the-Loop Validation**: While the system extracts the data automatically, the interface is designed for mandatory user review. The side-by-side layout ensures care coordinators can verify the AI's accuracy against the source document before committing the data to the system. * **Status Indicators**: Visual cues (like the green and red dots next to section headers) provide immediate feedback on data validation, highlighting sections that are complete, missing mandatory fields, or require manual attention. * **Care Coordination Interoperability**: By converting static PDFs into dynamic profiles, the system prepares the data for downstream workflows, such as scheduling, care delivery, and billing. By executing an automated, 13-step pipeline, this agent ingests dense NJ DDD care plans (ICP/NJISP) and Service Delivery Records (SDR). It automatically provisions a fully functional client profile, maps insurance billing, and configures caregiver tasks in under three minutes. The agent is trained to identify the following data items and inserts them into the Fieldworker.
### The Core Problem: The Administrative Tax on Care In 2026, social care agencies are trapped between a historic caregiver shortage and increasingly aggressive state and federal billing audits. State audit bodies are deploying automated tools to catch mismatched codes and over-authorization errors. When agencies rely on legacy software, their administrative staff spends hours manually copying and pasting lines, mapping codes, and fixing data entry typos. A single mistake can trigger a retroactive claim denial months later. Our ingestion agent acts as an automated compliance shield, catching discrepancies and setting up a clean framework before a single unit of care is even delivered. ``` [Dense State Care Plan] │ ▼ ┌───────────────┐ │ Ingestion │ ──> Executes 13-step verification pipeline │ Agent │ ──> Applies "Smart Sync" to authorization blocks └───────────────┘ │ ├───────────────────────────────────────┐ ▼ ▼ THE AGENCY IMPACT THE EMPLOYEE IMPACT · Zero-typo claim routing · No technical anxiety · Protected historical data · Automatic task-to-goal binding · Audit-ready transaction logs · Clear, visible client outcomes ``` ### User Guide The Data Ingestion & Onboarding interface (currently focused on the Upload SDR File module) serves as the primary entry point for converting static state and regional documents into actionable, structured data within the FieldWorker platform. By automating the parsing of critical onboarding documents, this agent establishes the foundational authorizations necessary for service providers submitting claims based on procedures that are already pre-authorized. ### 1. Interface Overview Located under the New module in the global sidebar, this section acts as a centralized creation hub. The interface is optimized for accuracy, compliance, and swift data entry. #### The Creation Toolbar (Top Right) A persistent horizontal menu allows users to rapidly switch between different creation and ingestion flows. Currently active is the SDR (Service Detail Report) upload, but the toolbar also provides one-click access to initialize a: * Task * Project * Contact * Customer * Note * Location * Bulk Upload * Activity * Intake * Workflow #### The Upload Dropzone The primary interactive area for document ingestion features a clear drag-and-drop target with strict pre-validation guardrails to ensure data integrity before the parsing agent runs: * Accepted Formats: Restricted exclusively to valid NJ Individualized Service Plan (NJISP) or Service Detail Report (SDR) PDFs. * File Constraints: Maximum file size of 5 MB, processed one file at a time to ensure accurate data mapping. * Security & Compliance Lock: The system cross-references the agency name within the document against the logged-in company profile (e.g., *Analytiq Inc*). Mismatches are automatically rejected, preventing cross-tenant data contamination.
#### Recent Uploads Ledger (Right Panel) This section acts as a real-time audit trail and processing queue for the ingestion agent. * Status Tracking: Displays the document name, identified type (e.g., SDR or NJISP), timestamp of upload, and the user who initiated it. * View Results: The VIEW RESULTS action allows staff to review, verify, and approve the structured data the agent successfully extracted from the unstructured PDF before it is permanently committed to the client's profile. ### 2. Core Workflows: Document Ingestion Proper utilization of this agent eliminates manual data entry, reducing human error when setting up complex client care plans and pre-authorized service limits. #### Step 1: Initiating an Upload 1. Navigate to the New module on the left sidebar. 2. Ensure SDR is highlighted in the top-right Creation Toolbar. 3. Review the *Before you upload* checklist to confirm your document meets the strict formatting and security criteria.
#### Step 2: Processing the Document 1. Drag and drop the target PDF (NJISP or SDR) into the upload box, or click to open the file explorer. 2. Once uploaded, the AI ingestion agent will begin reading the document, extracting vital information such as demographic data, approved service codes, authorized units, and effective dates.
#### Step 3: Verification and Activation (Human in the Loop) Post-processing, all expected changes are presented to the user for verification. At this point, human verification and explicit action are needed before the AI Model commits any changes to the system. #### Step 4: Review 1. Locate your processed file in the Recent uploads panel on the right. 2. Click VIEW RESULTS. 3. *Critical Action:* Review the extracted data against the original PDF. Once verified, this data populates the client's profile, automatically configuring the necessary guardrails for downstream EVV (Electronic Visit Verification), scheduling, and clean claim generation.
### 3. System Cross-Referencing The data extracted by this onboarding agent acts as the definitive source of truth for several downstream FieldWorker modules: * Projects & Clients: Automatically generates or updates the client's primary care profile and active project parameters. * Scheduler & Workflows: Ingested service authorizations dictate how many hours or units can be scheduled, preventing staff from over-servicing beyond approved limits. * Billing Engine: By mapping the exact pre-authorized procedures directly from the SDR/NJISP, the system ensures that all subsequent timesheets and EVV records are directly tied to valid, billable codes, virtually eliminating claim rejections due to unauthorized services. ### The Dual-Vector Value Proposition #### The Agency Perspective: Revenue Protection & Institutional Security For agency owners and directors, manual data ingestion is both an operational bottleneck and a financial risk. This agent systematically de-risks the onboarding lifecycle: * Eliminating Front-End Denials: The agent automatically pairs incoming document data with your agency's verified, claims-supported payers. It bypasses the manual text entry phase entirely, removing the typos that cause instant electronic claim rejections. * Non-Destructive Plan Updates: When an individual’s care plan is updated by the state, the agent uses intelligent merge logic. It overlays new authorization details without accidentally corrupting or wiping out the historical notes and custom fields you’ve built up over the years. * Locked Billing Infrastructure: The agent automatically extracts procedure codes, builds the required task types, and binds them explicitly to state-approved outcomes. This locks the system down so caregivers cannot accidentally log unbillable hours. * Audit-Ready Logs: Every single ingestion triggers a permanent, unalterable transaction record. If your agency faces a state audit years down the line, you have absolute proof-of-origin for how that billing workspace was generated. #### The Employee Perspective: Eradicating Burnout & Friction Direct Support Professionals (DSPs), case managers, and intake coordinators should be focusing on human lives, not wrestling with rigid software layouts. * Removing Tech Anxiety: Frontline staff shouldn't need a degree in data science to onboard a client. The agent handles deep-tech operations—like running background full-text token searches against master drug registries to log complex medications—completely behind the scenes. * Instant Clinical Clarity: Instead of forcing staff to hunt through a 40-page state PDF to find what they are supposed to be tracking, the agent instantly formats care plans into clean, localized goals categorized by outcome numbers. * Elimination of Repetitive Tasks: Employees skip hours of tedious drop-down configuration. Because the agent auto-generates and maps task types from the baseline authorization strings, the system handles the configuration work automatically. * Standardizing Data Across Teams: In a sector with high turnover and diverse technical literacies, the agent standardizes data ingestion. It ensures that every profile is built perfectly every time, regardless of the onboarding coordinator's software experience level. > #### The Bottom Line > > This ingestion agent transforms FieldWorker and AbilityHub from basic tools into defensive infrastructure. By automating the highest-friction touchpoint in operations, we reduce administrative overhead for agency owners while drastically cutting down the daily data-entry burden on employees. ### Human in the Loop: Reviewing and Confirming SDR Extractions This workflow occurs after a document has been uploaded via the New > SDR pathway.\ **Step 1: Navigating the Split-Screen View** * Left Panel (Source Document): The original uploaded PDF (e.g., `Analytiq_Raymond_ISP.pdf`) is displayed here. You can scroll through the document, zoom in/out, and rotate pages using the toolbar at the top of the PDF viewer to locate specific information. * Right Panel (Extracted Data): The system displays the parsed data categorized into logical sections (e.g., Customer Information, Diagnosis Information, Customer Goals, Service Authorizations).
**Step 2: Validating Extracted Information** * Review the status indicator next to each section in the navigation list (e.g., the green dot next to "Customer Information" indicates all required fields were successfully captured). * Read through the extracted values on the right and cross-reference them with the original document on the left to ensure accuracy (e.g., verifying the Medicaid ID, Service Units, and Authorization Dates).
**Step 3: Editing and Correcting Data** * If an extraction is incomplete, incorrect, or flagged with a red indicator (such as the "Medical Practitioners" section), click the EDIT button located at the top right of that specific section. * Manually enter or correct the missing information directly within the provided fields. * Save the edits to update the section's validation status.
**Step 4: Finalizing the Intake** * Once all sections have been reviewed and corrected, navigate to the top of the page. * Click PROCEED to commit the extracted data to the FieldWorker database, creating or updating the client's profile, support needs, and service authorizations. * If the document was uploaded in error or the extraction needs to be abandoned, click CANCEL.
#### III. Practical Use Cases **1. Rapid Intake for Specialized Populations** When onboarding individuals with intellectual and developmental disabilities (I/DD), agencies receive lengthy, comprehensive ISPs detailing extensive health, behavioral, and support needs. Instead of manually typing out diagnoses, emergency contacts, and daily care requirements (e.g., assistance with dressing, specialized diets, or mobility needs), the SDR instantly populates the client's electronic profile, saving hours of administrative labor and reducing transcription errors. \ **2. Streamlining Pre-Authorized Claim Preparation:** The SDR accurately captures complex Service Authorizations directly from the state document—including procedure codes (e.g., T2019HI), units, rates, frequency, and authorized dates. By digitizing this data accurately, the system establishes the foundational billing parameters required for service provider agencies submitting pre-authorized claims, ensuring that scheduled services align perfectly with approved budgets and billing constraints. \ **3. Updating Expiring Authorizations or Rate Changes:** When an updated ISP is issued due to a new plan year or a state-mandated rate change, the document can be processed through the SDR. The system will identify the new effective dates and rate adjustments, allowing the agency to seamlessly update the client's financial profile without manually auditing the entire 15+ page document.