Grant Proposal Intake and Pre-Review Workflow Spec

Role: Director of Research Operations and AI Workflow Engineering.

Context

• Research administration unit: {{university_research_office}}
• Funder regulatory guidelines: {{funder_guidelines}}
• Compliance protocol: {{ethics_board_protocol}}
• Evaluation rubric: {{scoring_rubric}}
• Triage turnaround SLA: {{triage_latency_target}}
• Human escalation condition: {{escalation_threshold}}

Task

Formulate a rigorous workflow chain specification for an automated grant pre-review pipeline that verifies administrative compliance, conducts ethical screening against {{ethics_board_protocol}}, and triages proposals across {{university_research_office}} within {{triage_latency_target}}.

Method

1. Specify the document ingestion and structural validation agent for proposals submitted under {{funder_guidelines}}.
2. Design the budgetary and administrative compliance auditing chain to flag disallowed costs, missing sub-awards, or formatting deviations.
3. Formulate the ethics pre-screening agent evaluating human subject, animal care, and export control risks against {{ethics_board_protocol}}.
4. Construct the semantic profile matching agent that parses project abstracts to identify appropriate faculty reviewers using institutional bibliographies.
5. Establish the rubric scoring agent that populates preliminary readiness indices using {{scoring_rubric}}.
6. Detail the automated triage router directing proposals to either standard fast-track approval or {{escalation_threshold}} escalation paths.
7. Specify notification and summary packet generation for research deans and principal investigators within {{triage_latency_target}}.

Constraints

• MUST flag non-compliant grant applications before any qualitative peer review matching occurs.
• MUST NOT provide subjective scientific merit judgments; agents must restrict evaluation strictly to {{scoring_rubric}} compliance.
• Reviewer matching must exclude any faculty member with identified institutional or co-authorship conflicts of interest.
• Pipeline execution must complete end-to-end processing within {{triage_latency_target}}.

Output format

1. Workflow Architecture & Stage Map (Numbered sequence with stage gates)
2. Agent Execution Specifications (Agent Name, Task, Input Schema, Output Schema, Validation Rules)
3. Compliance & Ethics Matrix (Rule mappings for {{funder_guidelines}} and {{ethics_board_protocol}})
4. Reviewer Matching & Conflict Algorithm Specification (Logic criteria)
5. Escalation & Triage Policy Document (Structured rules for {{escalation_threshold}})

Self-review

• Ensure conflict-of-interest checks are explicitly detailed in the matching stage.
• Verify that administrative compliance verification strictly enforces {{funder_guidelines}}.
• Confirm that escalation workflows correctly trigger when metrics meet {{escalation_threshold}}.