How Grant Hawk works
Operating entirely within the University of Iowa ChatGPT Edu environment, Grant Hawk generates reviewer-style critiques using patterns derived from real grant applications, summary statements, and literature central to a proposal’s scientific premise. It produces detailed narrative reviews with three main components:
1. Three-perspective panel review
The review is designed to reflect realistic panel dynamics, reviewer disagreement, and NIH-style decision-making. Like a real NIH study section evaluation, each proposal is reviewed independently by three distinct reviewer perspectives. Each reviewer emphasizes different aspects of the application such as innovation, methodological rigor, feasibility, translational impact, or mechanistic interpretation. It also generates a simulated panel discussion that captures simulated reviewer agreement, disagreement, shifts in enthusiasm, and the dominant issues most likely to shape competitiveness and funding potential.
2. Review of NIH criteria
The review includes an overall impact assessment, reviewer-style critiques organized by the traditional NIH review criteria: significance, innovation, approach, investigator(s), and environment. As we gather more input data in the database over time, we will transition feedback to the new NIH review format.
3. Suggestions to improve competitiveness
Grant Hawk provides a structured analysis of the proposal’s major competitiveness drivers, identifies the concerns most likely to suppress reviewer confidence, and highlights the highest-leverage revisions most likely to improve funding potential.
Important limitations
Grant Hawk should supplement, not replace, your judgment, program officer guidance, mentor feedback, and/or human scientific review. It is based in AI and simulates reviewer reasoning and NIH-style critique patterns, but it does not predict funding outcomes or guarantee application success.
Outputs should be interpreted as structured feedback intended to improve proposal clarity, competitiveness, and reviewer understanding prior to submission. Misguided or inaccurate guidance may occasionally be offered.
Get started
Eligibility
You may submit your application for review if:
- You are University of Iowa faculty
- You are the PI or a Multiple Principal Investigator (MPI) of the application
- Your collaborators agree to the submission
- Your grant application is near completion
- Your grant is due to NIH in four weeks or more
July 2026 submissions will receive priority and may be submitted three weeks prior to NIH submission deadlines.
Submission requirements
To generate a NIH panel simulation, provide materials that would shape real reviewer evaluation in addition to the grant proposal package. The review quality depends heavily on having the same scientific, strategic, and contextual information available to the AI system that NIH study section members would know or use.
1. Create a OneDrive folder
2. Share "edit" access with the following RDO team members: Aaron Kline, Britt Ryan, and Jill Wiley.
3. Upload the following materials to the OneDrive folder:
- Grant application PDF
- If applicable, resubmission materials such as the PDF, A0 application, and summary statements
- Prior reviewer feedback
- Approximately 15 significant journal articles directly related to the proposed science such as:
Articles that support the premise of the proposal
Articles presenting competing models
Articles offering alternative interpretations
Articles representing differing perspectives within the field
4. Fill out the submission form
Do NOT submit protected health information, export-controlled information, proprietary industry materials without authorization, or other restricted or confidential information that is not appropriate for institutional AI systems.
Please note: your grant proposal and submission materials may be used to improve Grant Hawk for future users. All materials and data will stay within the University of Iowa's ChatGPT Edu workspace, which has enterprise data protection.
You’ll receive your review in under five business days.
Frequently asked questions
Who should use Grant Hawk?
Grant Hawk is designed for UI faculty preparing NIH submissions who desire a rigorous pre-submission critique that approximates study-section reasoning and panel dynamics. It is particularly useful for applications that are near completion, resubmissions responding to prior reviewer critiques, complex interdisciplinary proposals, and applications where researchers want additional perspective on competitiveness, clarity, or reviewer risk perception before submission.
How soon should I submit my application for review?
Submit your grant application for review at least four weeks before the agency deadline. RDO will complete the review in less than five business days, giving you time to evaluate and incorporate feedback prior to internal deadlines and submission to NIH.
For those with grants due in July 2026, the team will accept proposals due within three weeks and will prioritize the review.
How do I choose which 15 journal articles to submit?
To provide a realistic review, select articles that are directly related to the proposed science and reflect the range of perspectives and scientific discussion in the field. This includes articles that:
Are in your reference section as well as those that are not.
Support your central premise as well as alternative interpretations, competing models, or differing perspectives within the field.
Are in top-tier journals for the field.
May be familiar to NIH study section members.
Theoretical, empirical, and review papers are all appropriate.
Do you provide reviews for agencies other than NIH?
Not yet, but we are working to refine versions of Grant Hawk tailored to other federal agencies, like the National Science Foundation (NSF). If you are interested in participating in a future pilot project for NSF or other agencies, contact RDO at vpr-rdo@uiowa.edu.
How did RDO develop Grant Hawk?
The RDO team developed Grant Hawk within the University of Iowa ChatGPT Edu environment and trained it on publicly available NIH grant proposals and feedback, both funded and unfunded. Faculty members representing multiple colleges participated in the pilot phase, provided materials with permission, and shared feedback on the outputs. Overall, faculty feedback was positive and constructive. For example, Juliana Souza-Talarico, associate professor in the College of Nursing, said: “It felt so realistic that, as I was reading, I could clearly picture an NIH reviewer in my mind and feelings, while reacting to specific sentences. It’s excellent and directly addressed a concern I had been carrying, but that hadn’t been raised by my co‑investigators or external reviewers.”
Will you issue a final version of Grant Hawk?
The RDO team is continually improving Grant Hawk. They may contact researchers who use it to gather review comments and summary statements to refine the feedback provided.
For additional questions or support, contact RDO at vpr-rdo@uiowa.edu.