AI / Automation Workflow Stack (OMB M-25-21 / M-25-22 Compliance)
RD-FO has identified — and OMB M-25-21 / M-25-22 mandate — that AI / automation should accelerate input preparation, anomaly flagging, certification-sheet generation, and similar mechanical work across the credit-reform, accounting, and budget-execution workstreams, but RD does not yet have an AI workflow stack that meets federal AI governance requirements.
Why It Matters
OMB M-25-21 and M-25-22 are not optional — they are the framework under which federal agencies must adopt AI. Standing up a production-ready AI workflow stack is a foundational OBP capability for the next 24 months.
HSG's Approach
- 1Deploy HSG's AI workflow methodology as the OBP AI overlay, with deployment on Anthropic Claude API or Azure OpenAI in federal-cloud-eligible regions.
- 2Build the four flagship AI workflows: (1) AI-augmented ULO triage; (2) Re-estimate memo auto-draft; (3) Burn Rate / FTE forecasting scenario tool; (4) Anomaly detection at $40M / 5% threshold.
- 3Maintain the AI / Expert Reconciliation Log as the signature transparency deliverable for OIG, GAO, and A-123 audit defense.
- 4Align governance to OMB M-25-21, M-25-22, EO 14179, and NIST AI Risk Management Framework, with documented model cards, use-case registration, and human-on-the-loop discipline.
- 5Train OBP and NFAOC analysts on each workflow with stand-up SOPs and an 'RD-maintainable' tier.
Expected Deliverables
- AI workflow deployment in federal-cloud-eligible region
- Four flagship AI workflows in production
- AI / Expert Reconciliation Log
- AI governance package aligned to OMB M-25-21 / M-25-22
- Analyst training and 'RD-maintainable' SOPs
Expected Outcome
Establish an OMB-M-25-21-compliant AI workflow stack delivering measurable cycle-time and quality improvements across the four flagship use cases within 6 months of contract start.