EVP Artificial Intelligence Innovation Program (AIIP)

EVP Artificial Intelligence Innovation Program (AIIP)

Introduction

The Executive Vice President (EVP) Artificial Intelligence Innovation Program (AIIP) is intended to encourage the exploration of generative AI to improve operational efficiency and administrative systems and processes. The AIIP is an inclusive approach to innovation—soliciting proposals directly from EVP staff—and is designed to foster creativity and collaboration among our respective Units’ teams.

Program Details

For our inaugural round of submissions, staff were invited to submit proposals to address a business problem or opportunity with AI. Proposals were reviewed by a committee with representation from across EVP areas. Selected proposal teams receive:

  • Access to end-user AI tools such as ChatGPT Enterprise, Microsoft CoPilot, and Google Duet and AI services including AWS Bedrock, Azure OpenAI, and Google Vertex.
  • Support from HUIT and strategic technology partners, including Amazon AWS, to explore AI solutions, lend expertise, and help shape your proposal.
  • Up to $15,000 in funding for a 6 month experiment or pilot.
  • An opportunity to showcase key learnings and AI innovations to administrative leadership.

Suggested proposals included, but were not limited to, the following:

  • Administrative efficiences and productivity gains, e.g., scheduling and calendar management, etc.
  • Customer service improvements, e.g. chatbots in customer-facing websites
  • Accessing information more quickly, e.g., parsing long data sets
  • System improvements, e.g., incorporating AI into our existing tools
  • Workflow automation, e.g., invitation letters and confirmation of receipt, automating website updates, etc.
  • Financial operations / procurement improvements, e.g. Onboarding, consolidating, standardizing, identifying suppliers or spend, credit card management and analytics, eliminating invoice exceptions, budgeting support

Below are summaries of EVP AIIP project proposals that have entered the pilot stage. 

First round pilots

AI for Administrative Learning Resources 

 

  • Leverage AI to create draft training content for sponsored research topics at Harvard, significantly reducing SMEs' development time. 
  • Increase efficiency, accelerating content delivery while reducing non-compliance risks and audit exposure. 

AI-TIES (Targeted Information Extraction and Summarization) 

 

  • Use generative AI to automate information extraction and create task-specific summaries, improving efficiency and demonstrating AI's ability to filter pertinent content. 
  • Aim for precise information extraction and summarization, lessening manual distillation effort. 

HUHS FirstAIde 

 

  • Develop an AI chatbot for Harvard University Health Services (HUHS) customer support staff, offering digital health plan services and integrating AI for scaled customer service and improved response times. 
  • Improve customer service by providing faster, precise responses and increasing operational efficiencies, serving as a learning tool for Member Services staff. 

Improving Reunion Customer Service with AI 

 

  • Employ AI tools for email review and triage in Alumni Affairs & Development areas across campus, streamlining large volume email processing. 
  • Save staff time, enhance efficiency in handling requests and feedback, and provide effective support in donor-related tasks. 

Productivity, Risk-Mitigation, and Efficiency through AI-Assisted Contract Drafting 

 

  • Utilize AI for contract drafting and analysis at the Office of Contract Management (OCM), enabling faster and more reliable self-generated drafts and risk assessments. 
  • Enhance OCM’s efficiency and effectiveness in contract review and negotiation, reducing risks and achieving savings. 

ServiceRightNow 

 

  • Use AI, particularly ChatGPT, to automate routine service ticket tasks, focusing human expertise on complex issues. 
  • Expect a significant reduction in workload, assessed through fewer tickets, quicker response times, and improved customer satisfaction. 

Use of AI in Streamlining Data Use Agreements​ 

 

  • Apply AI to streamline processing of Data Use Agreements (DUAs), focusing on identifying standard language exceptions to expedite the review process. 
  • Realize a substantial decrease in initial review time, easing the burden on reviewers and negotiators handling complex negotiations. 

Using AI to Enhance the HR Policy Creation and Review Processes​ 

 

  • Implement AI to analyze data and improve HR policy creation and review at Harvard, focusing on searchability, accuracy, consistency, relevance, accessibility, or inclusivity. 
  • Result in a more accessible, organized policy library, facilitating strategic focus for workforce policy leadership.