X

5th AI Symposium Sessions

Education Fifth AI Symposium 5th AI Symposium Sessions

Event Schedule

Time Session Title & Presenter(s) Track & Session Description Learning Objectives Prerequisites
9:00-10:15 AM Keynote Come back later to see who will be presenting the opening keynote of NCURA's 5th AI Symposium!    
10:30-11:45 AM Operationalizing Responsible AI: Governance, Risk, and Compliance in Research Administration - Presented by: Kathleen Halley-Octa, Attain Partners & Gayle Sherwood, California State University-Monterey Bay Track: AI Governance and Responsible Adoption
As artificial intelligence becomes increasingly embedded in research administration, institutions face growing pressure to ensure its use aligns with regulatory requirements, ethical standards, and institutional risk tolerance. The challenge is no longer simply adopting AI tools but operationalizing responsible AI in environments shaped by federal funding rules, data governance constraints, and evolving compliance expectations. This session provides a practical framework for designing and implementing AI governance in research administration. Attendees will explore how to identify and mitigate risks related to data privacy, security, bias, auditability, and appropriate use, while enabling innovation and operational efficiency. We’ll examine how to align AI use with existing compliance structures, including research security, human subjects protections, financial oversight, and proposal development practices. Grounded in real-world scenarios, the session will also address how to establish policies, review processes, and monitoring mechanisms that scale with institutional adoption. Participants will leave with actionable approaches to embedding responsible AI practices into their organizations without slowing progress or overburdening teams.		 In this session, participants will:
1. Establish a Governance Framework for Responsible AI - Design governance structures that define roles, responsibilities, and decision rights for AI use in research administration, aligned with institutional policies and regulatory expectations.
2. Identify and Mitigate AI-Related Risks Across the Research Lifecycle - Assess key risk domains—including data privacy, security, bias, compliance, authorship, and auditability—and apply practical strategies to mitigate them across activities such as proposal development and human subjects research.
3. Integrate AI Oversight into Proposal Development, and Compliance Processes - Align AI governance with established institutional frameworks (e.g., IRB review, proposal development, research compliance, IT security, and data governance) to enable scalable oversight, appropriate disclosure, and sustainable adoption.		 Basic
10:30-11:45 AM From Manual to Magical: Using Copilot to Build a CPOS Tracker-to-XML Generator for SciENcv - Presented by: Rochelle Hubbart, University of Colorado Anschutz Medical Campus & Holly Heilman, University of Colorado Anschutz Track: AI in the Research Administration
Preparing Current and Pending Other Support (CPOS) documents for SciENcv can be a time-consuming and error-prone process—particularly when data is entered manually. This session demonstrates how Microsoft 365 Copilot can be leveraged as a tool development partner.  Attendees will be guided through the creation of an Excel-based CPOS tracking template that transforms structured data into XML and integrates with a Power Automate workflow to generate SciENcv-compatible XML files ready for upload. The session will cover AI ethics, effective prompt strategies for working with Copilot, key design considerations for structuring CPOS data, and practical approaches for connecting Excel outputs to automated workflows.		 In this session, participants will:
1. Identify key ethical considerations when incorporating AI tools into research administration processes, such as Other Support reporting.
2. Explain prompt engineering techniques to guide Copilot in generating formulas, logic, automation components, and conditional formatting.
3. Learn how to design a structured Excel-based CPOS tracking template that supports conversion into XML format for SciENcv.
4. Explain how Excel outputs can be integrated with Power Automate to generate SciENcv-compatible XML files.		 Advanced: Familiarity with Current & Pending (Other) Support document
10:30-11:45 AM Should I Really Trust AI? Navigating Risk, Responsibility, and Real-World Use in Research Administration - Presented by: Aaron Wolpoff, Streamlyne & his institution partner Track: AI Tools, Demonstrations, and Case Studies
AI is already part of a modern research office, sometimes in invisible ways. It’s accelerating workflows, supporting decision-making, and helping teams do more with limited resources. But as adoption increases, so do critical questions:
• How much should you trust AI?
• What’s actually happening with your data?
• Where are the real risks, and where are they overstated?
• How are peer institutions using AI in ways that are both effective and responsible?
Drawing on more than seven years of experience building AI specifically for research administration, we’ll provide a clear view into how responsible AI systems are designed, the safeguards that should be in place, and how to distinguish between tools built for institutional use and those that may introduce unnecessary risk.
You’ll leave with a practical framework for evaluating AI solutions, guidance on engaging with AI tools in ways that improve outcomes without increasing exposure, and a stronger sense of how to lead A		 In this session, participants will:
Assess when and how to trust AI in research administration environments; Identify risks and safeguards across different types of AI tools; Apply best practices for using AI securely and effectively		 Basic
12:45-1:15 PM Sponsored Demos Come back later to read about our sponsored demos at NCURA's 5th AI symposium!    
2:15–3:30 PM Demonstrating AI-Driven Award Data Extraction, Reconciliation, and Setup - Presented by: Chris Steele, Huron & Mary Catherine Gaisbauer, University of California - Santa Barbara Track: AI Tools, Demonstrations, and Case Studies
Award setup is still largely a manual, document-driven process, whether institutions are in the middle of a system conversion or just trying to keep up with daily intake. In this session, we will walk through an AI-enabled award extraction tool that the University of California Santa Barbara used during its Oracle implementation. This will include the process used and what can be done with current solutions to support successful conversions. We will also focus on how the new tool can be used with day-to-day award setup. The session will focus on showing how the tool actually works in practice. We will demonstrate how research administrators use it to upload award documents, extract key award demographics and terms, and review the results directly alongside the source language. The goal is not to eliminate human review, but to reduce re-keying, speed up validation, and surface issues earlier in the award setup process. We will also show how the same tool was used during UCSB’s Oracle conversation.		 In this session, participants will:
1. See how an AI award setup tool is used as part of normal, day-to-day research administration workflows.
2. Understand how AI-extracted award data can be reviewed and validated using clear source context.
3. Learn how the same tool can be applied to data reconciliation during a system conversion.
4. Take away practical lessons from UCSB’s experience that can be applied to award setup and data quality efforts at other institutions.		 Basic
2:15–3:30 PM Reducing Pre-Award Burden with AI: Smarter Matching, Stronger Proposals, Better Outcomes - Presented by: Christine Cline, Auburn University, Rob Ellis & Steve Pinchotti from Altum Track: AI in the Research Administration
Research administrators are under increasing pressure to support more faculty, manage complex sponsor requirements, and handle time-intensive tasks like opportunity searches and eligibility checks, all while delivering high-quality proposals with limited resources. Artificial intelligence is emerging as a practical way to reduce administrative burden. This session explores how AI can be integrated into core research administration workflows, with a focus on funding opportunity discovery, eligibility assessment, and proposal development. Attendees will learn how AI-powered tools can help match faculty and institutions to the most relevant opportunities, automate initial eligibility checks against sponsor requirements, and generate proposal content aligned to specific RFP requirements. The session will also highlight how these capabilities connect to broader innovations across the grants ecosystem, including AI-supported peer review analysis, overlap detection, and conversational tools, demonstrating a more unified, data-driven approach to research administration. Participants will gain practical insight into how platforms, such as ProposalCentral.ai and Altum Intelligence, are enabling institutions to better support faculty and accelerate the path from opportunity identification to submission. The discussion will also address responsible implementation, including strategies for safeguarding proprietary data and ensuring AI tools are used in alignment with institutional policies and sponsor expectations.		 In this session, participants will:
1. Identify common pre-award pain points (e.g., manual opportunity searches, repetitive eligibility checks, early-stage proposal drafting) and explain how AI can reduce administrative burden 2. Understand how AI-driven tools can improve funding opportunity discovery and eligibility matching for faculty and institutions 3. Describe how AI can support proposal development with tailored guidance aligned to sponsor requirements 4. Recognize how pre-award AI capabilities connect to broader innovations across the grants lifecycle, including review and compliance support 5. Evaluate key considerations for implementing AI solutions in a research administration setting, including data security, governance, and responsible use		 Intermediate: Participants should have a basic understanding of general grant and research administration workflows (i.e., pre-award processes, proposal development, or sponsored programs support)
2:15–3:30 PM What Research Administrators Actually Think About AI — And Why Leadership Should Listen - Presented by: Dan Harmon, University of Illinois at Urbana-Champaign & Nihal Sarikaya, Northern Arizona University Track: Future of Research Administration in the AI Era
Artificial intelligence is reshaping the landscape of research administration, but how well do institutional leaders understand what's happening on the ground? This session presents findings from multiple surveys conducted across the research administration community, synthesized and analyzed by the REACH AI Working Group — a collaborative body focused on data, analytics, and evaluation for research-focused institutions. Together, these datasets offer a rare, multi-institutional window into how research administrators are experiencing AI adoption firsthand: what they're using, what they're worried about, and where they feel unsupported. Rather than a top-down technology briefing, this session centers the voices of practitioners. Attendees will explore where meaningful gaps exist between administrator perspectives and institutional leadership priorities — across dimensions like governance, training investment, workload impact, and trust in AI outputs. REACH is supported by the U.S. National Science Foundation (grant number: 244978). Content of this session do not reflect the opinions or policies of the NSF.		 In this session, participants will:
1. Identify key themes and trends in AI adoption and sentiment among research administrators, drawn from multi-institutional survey data collected through the REACH AI Working Group.
2. Recognize common gaps between front line administrator experiences and institutional leadership priorities around AI governance, training, and implementation.
3. Apply a practical framework for assessing their own institution's alignment between staff-level AI experiences and leadership strategy.
4. Articulate evidence-based recommendations to leadership and stakeholders that reflect the workforce perspective on AI readiness and support needs.		 Basic
3:45–5:00 PM AI4RA - Finding an Actionable On-Ramp for Implementing AI Solutions in Research Administration - Presented by: Parker Grimes, Southern Utah University, Nathan Layman, University of Idaho & Nathan Wiggins, Southern Utah University Track: AI Tools, Demonstrations, and Case Studies & Track: AI Skills for Research Administrators (Hands On)
This session is an interactive workshop, designed to equip Research Administrators with actionable strategies for implementing AI tools at their own institution. Participants will gain hands-on experience testing free open-source tools developed by the AI4RA team that they can readily adapt and deploy themselves, along with an overview on configuring a cost-effective server to host local AI models. Recognizing the value and challenges of working with IT teams, the session will also offer insight into securing approval, building support, and navigating internal processes for AI initiatives. Whether participants are actively deploying AI solutions or are still trying to find the right on-ramp, this immersive learning experience is designed to provide value that meets research administrators where they are - emphasizing active participation and plenty of hands-on sandbox time to experiment with AI tools, server setups, and implementation approaches.		 In this session, participants will:
1. Explore the intersection of AI and data science, including gaining access to open-access AI tools that have been developed specifically for research administration use-cases.
2. Learn how to set up a server for hosting local AI models and gain familiarity with server infrastructure.
3. Inherit strategies for coordinating with your institution's IT department to gain approvals and support in implementing AI solutions.		 Intermediate: None, Participants should plan on bringing a fully charged laptop to this session.
3:45–5:00 PM Ask Anything About Your Research Data: A GenAI Solution for Natural Language Queries on Tabular Datasets - Presented by: Chandni Mathur, University of Illinois Urbana-Champaign Track: AI Tools, Demonstrations, and Case Studies
Research administrators and institutional leaders frequently need timely insights into proposal volumes, department- and PI-level distributions, grant award trends, and detailed expenditure patterns across departments, PIs, and research areas. Traditional approaches rely on manual analysis, custom reports, or static PowerBI/Tableau dashboards. While effective for predefined metrics, these methods fall short when addressing ad-hoc or unexpected questions that arise during strategic discussions. This presentation introduces a practical Generative AI (GenAI) solution that functions as an intelligent, on-demand research data analyst. The chatbot has secure access to raw proposals, awards, and expenditure datasets. Using a multi-agent architecture, it intelligently routes queries to specialized agents (e.g., Proposal Agent, Awards Agent) or coordinates multiple agents as needed. For each question, the system dynamically generates and executes Python code to retrieve accurate answers directly from the source data. Unlike static reports or dashboards, this tool enables users to ask any natural language question in real time, delivering immediate, contextual responses without waiting for analysts or report generation. This significantly accelerates decision-making and empowers leadership with flexible data exploration. The session will include a live demonstration of the tool along with key implementation considerations such as data security, accuracy safeguards, and responsible AI governance.
Example queries include:
1. “Which department has the highest rate of grant awards?”
2. “Which agency is our institution’s highest sponsor?”
3. “Compare proposed versus actual expenditures by sponsor for the last fiscal year.”		 In this session, participants will:
1. Understand how a Generative AI (GenAI) multi-agent solution can transform access to research administration data by enabling natural language queries on tabular datasets (proposals, awards, and expenditures).
2. Explore the architecture and workflow of the tool, including specialized agents and dynamic Python code generation for accurate query responses.
3. Learn how to adapt and apply this GenAI framework to your own institutional use cases, data sources, and research administration needs.
4. Evaluate critical implementation considerations including data security, accuracy safeguards, and responsible AI governance.		 Intermediate: It is recommended but not required that attendees have some familiarity with programming languages, technical concepts, and a basic understanding of Large Language Models (LLMs).
3:45–5:00 PM Come back later to read about the 3rd breakout session at this time slot!