Procurement notices, grants, and career pages read for the underlying data or automation problem — not the job title — then analyzed for what an agent-run engagement could do about it. Everything the organization said is quoted; everything we inferred is labeled ours. These are not bounties. Nobody has escrowed anything. They’re where bounties come from.
Strategic Weapons Facility Pacific (SWFPAC) operates a mission-critical Microsoft Access 2003 legacy database for warehouse and inventory management (tracking locations, personnel, and two inventory categories). They need a complete migration to Microsoft 365 (Flank Speed), including a new SharePoint Online data architecture, Power Apps interface, and Power Automate workflows to replace the legacy system with minimal operational disruption.
An AI/data agent could automate Access database schema analysis, propose optimized SharePoint Online data models, generate synthetic test data aligned to the four core entities, validate relational integrity across warehouse locations, personnel, and dual inventory types, and enforce lifecycle management (check-in/check-out state transitions, archival rules). The agent would reduce discovery and architecture design cycles while ensuring compliance with audit and permission controls.
HHS ASA needs to develop and implement a federally aligned quality standard for healthcare data interoperability (USCDI Plus Quality) by extending the QI-Core standard to align with CMS quality measures and other requirements. The work includes standards governance, lifecycle management, stakeholder coordination, HL7 FHIR Implementation Guide development/testing/publication, and implementation support—all under HL7's exclusive governance authority.
An AI/data agent could automate standards lifecycle management (change tracking, impact analysis, versioning), validate Implementation Guide conformance against CMS measures through rule engines, manage stakeholder feedback and balloting workflows, and generate compliance documentation—accelerating standards development cycles while reducing manual coordination overhead.
The NIH needs to establish a centralized Data Coordinating Center that ingests, integrates, and makes accessible multiscale mouse brain datasets (multi-omics and connectomics) from the AD-MAP research projects. They have emerging datasets but lack unified infrastructure for harmonization, collaborative analysis, and public dissemination through a web interface.
An intelligent agent that orchestrates the ingestion, validation, harmonization, and dissemination of heterogeneous mouse brain datasets from contributing AD-MAP projects. It would automate metadata extraction, standardize data formats, detect and flag quality issues, suggest dataset alignments for cross-project analysis, and maintain the web-based visualization portal.
USDOJ-OJP-BJS has criminal history records from multiple state and federal repositories with inconsistent, jurisdiction-specific field definitions and text values. They need to convert these heterogeneous datasets into standardized, machine-readable formats suitable for statistical analysis of recidivism and criminal careers, while documenting data quality issues and harmonizing offense classification across jurisdictions.
An AI/data agent could automate the parsing of unstructured rap sheet text fields, apply and refine existing crosswalk tables to map state/federal offense codes to BJS standardized values, detect data quality anomalies (missing fields, inconsistent formats, jurisdictional variations), and generate transformation rules and documentation—reducing manual review overhead and accelerating the conversion of raw administrative records into analysis-ready datasets.
HHS-NIH needs a Research Coordinating Center to manage and harmonize multi-site clinical trial data and biospecimens across a consortium testing anti-obesity medications in youth. The center must establish common data elements, centralized laboratory procedures, database development, and cross-site data analytics while fostering research collaboration.
An AI/data agent could automate the ingestion, validation, and harmonization of data from multiple clinical sites using common data elements; manage laboratory and biospecimen workflows; flag protocol deviations and data quality issues; and generate standardized reports for cross-site analysis and collaboration.
USDOJ-BJS must collect, validate, and analyze criminal justice supervision data from a highly fragmented universe of 829+ state and local probation/parole agencies using inconsistent data systems, record formats, and reporting standards. They need to reconcile disparate data definitions (case counts vs. individual counts), aggregate surveys across three distinct instruments (felony parole, felony probation, misdemeanor probation), and produce accurate national correctional population estimates for 2026–2030.
An AI/data agent could automate reconciliation of disparate agency reporting formats, detect and flag data quality anomalies (e.g., case counts vs. individual counts mismatches), apply standardized validation rules across 829+ respondents, and generate reconciliation reports for human review before final aggregation and analysis.
NIH HEAL Initiative needs a Data Coordination and Integration Center to ingest, manage, and integrate complex pain-associated datasets from the INTERACT Consortium and related musculoskeletal research programs, then make them accessible through a web system for the research community. The center must handle multiple data types and progressively expand to integrate broader datasets across programs.
An autonomous agent that ingests heterogeneous clinical datasets from INTERACT and related musculoskeletal programs, validates and normalizes them to common schemas, manages metadata and data lineage, and exposes cleaned datasets through a web portal API with built-in discovery and hypothesis-generation features.
NSF seeks to fund collaborative cyberinfrastructure and data management projects that improve accessibility, discoverability, reliability, reproducibility, and sustainability of research data products across scientific domains. Organizations need to build competencies, capabilities, and communities around open science practices and data management infrastructure to align with federal public access and open science mandates.
An AI agent that automates metadata harvesting, cross-repository discoverability indexing, and compliance validation across federated research data repositories to ensure alignment with FAIR (Findability, Accessibility, Interoperability, Reusability) principles and NSF public access requirements.
HHS-NIH is funding a coordination hub to integrate data, evaluation, and community engagement across multiple CBPR oral health projects. The hub must provide technical assistance on research methodology and analytics, evaluate community engagement outcomes, manage consortium infrastructure, and enable data sharing across a new platform (SCHARE) while coordinating with external consultative resources (CEACR).
An AI agent could automate intake of data from multiple CBPR projects, standardize research methodology documentation, route technical assistance requests to appropriate experts, track and synthesize community engagement evaluation metrics, and feed aggregated insights to the SCHARE platform—reducing manual coordination overhead and improving data consistency across the consortium.
HHS-NIH needs to increase research productivity and discovery by making a large longitudinal dataset (71,000+ participants) on child health outcomes accessible and usable by postdoctoral fellows. The ECHO cohort's de-identified data is stored in DASH but requires analytical support to maximize its scientific value across exposure and outcome domains.
An AI agent could accelerate fellow productivity by automating data discovery (matching research questions to relevant ECHO variables), harmonizing multi-source exposure and outcome variables, and generating standardized analysis templates for common research designs across the five outcome areas.
NSF seeks to fund research and infrastructure projects that address FAIR (Findable, Accessible, Interoperable, Reusable) research data management, open science capabilities, data portals, research commons, and barriers to data access, curation, and sharing across scientific disciplines. Organizations must propose either disciplinary-specific or cross-cutting improvements to cyberinfrastructure, research data management practices, or scientific communication models.
An agent that assists funded research teams in automatically ingesting, cataloging, and linking research datasets across disciplinary domains using FAIR principles, standardizing metadata, generating discovery indices, and establishing interoperability between data portals and research commons platforms.
potl