DRAFT: Project Charter: Library Reference Chatbot

Project Goals

Increase self-service efficiency for common and repeatable library reference questions.
Improve user experience by providing 24/7 access to reliable reference support.
Seamlessly route complex or sensitive questions to appropriate library staff.
Reduce staff time spent on routine inquiries while preserving high-quality human support.

Problem and Value Statements

Problem Statement

Library users frequently have general reference questions that require staff mediation even though some answers may be routine or easily standardized. This creates delays for users and increases staff workload.

Solution and Business Value

Implement a chatbot designed to answer general library reference questions and escalate complex inquiries to staff.

Enables vendor-compliant traceability and auditability.
Aligns with Harvard’s Zero Trust and modern identity management principles.
Reduces technical debt and administrative overhead.
Enhances user equity and experience across all campuses and affiliations.

Alignment with Harvard Library Multi-Year Goals and Objectives

FY26 Objective 5: Advance the development and delivery of a range of consultation services and resources supporting the discovery, exploration, creation, and sharing of research and scholarship

Alignment with HUIT Objectives

Supports modern, scalable digital service platforms.
Leverages automation responsibly.

Vision

Library Reference Chatbot
A trusted virtual reference assistant that answers common questions and connects users to staff when needed.
Strategic Objectives	Guiding Principles	Key Performance Indicators
Provide accurate, consistent responses to common library reference questions through self-service. Improve user access to library information by offering 24/7 virtual assistance. Reduce staff workload for routine inquiries while preserving high-quality human support for complex needs. Ensure seamless escalation from chatbot interactions to appropriate library staff and service channels. Increase discoverability and effective use of library services, policies, and resources. Establish sustainable content governance to keep chatbot knowledge current and reliable. Collect and analyze usage data to inform service improvements and identify unmet user needs.	Implement AI-assisted services in alignment with library values of privacy, accessibility, and trust. Design the chatbot to prioritize clarity, accessibility, and ease of use for a diverse library user community. Provide responses that are grounded in approved, authoritative library information, with clear acknowledgment of uncertainty when applicable. Use automation to augment—not replace—library staff, with clear and reliable pathways to human assistance. Make it clear to users when they are interacting with an automated service and how responses are generated and escalated. Ensure the chatbot meets accessibility standards and supports equitable access to library services across time, location, and ability. Minimize data collection, protect user privacy, and comply with institutional and legal requirements for data governance. Establish clear ownership, governance, and processes to keep content current and the service reliable over time. Use analytics, feedback, and staff expertise to iteratively improve chatbot performance and service quality. Chatbot will work in concert with existing reference staff, and staff will develop a general understanding of its functions.	Increase successful self-service resolution of common reference questions Percentage of total chatbot interactions resolved without staff intervention Number of repeat questions successfully answered by the chatbot Improve user satisfaction and confidence in library reference services Post-interaction user satisfaction score User-reported clarity and usefulness of responses Reduce staff time spent on routine reference inquiries Reduction in volume of basic reference questions handled by staff Ensure effective and timely escalation to library staff when needed Successful handoff rate from chatbot to staff-supported channels Average time from escalation to staff response Support continuous improvement through analytics and feedback Frequency of knowledge base updates informed by analytics Reduction in unanswered or low-confidence chatbot responses over time Ensure responsible and ethical use of AI technologies Compliance with privacy and accessibility standards

In Scope/Out of Scope

In Scope

General library reference questions.
Integration with library websites.
Escalation workflows.

Out of Scope

In-depth subject research consultations.

Deliverables and Work Products

Task	Outcome	Responsible Parties
Functional Requirements Drafted	Clear definition of scope of included data sources and day one functionality outlined	LTS / HL
Vendor engagement to expose needed data.	Development agreements and schedule finalized.	LTS/Vendors
Communication campaign	Staff and faculty informed by Winter 2027	HL Communications / LTS / Project Team
General Availability of chatbot	Tool released for general use Spring 2027	HL / LTS

Definition of Done

Chatbot live in production.
Approved knowledge base.
Functional staff escalation.

Stakeholders

Stakeholder	Title/Role	Participation
VPDR	Project Sponsor	Approval and oversight
LTS	Lead implementer	Project management, system integration
HL Reference Team	Co-lead	Requirements gathering; testing
External Vendors	Development Partners	Creation of MCP layers, chat widgets
HL Communications	Support	Communication and messaging development
Public Services Staff	Operational partner	Guest support and user guidance

Project Team

Portfolio	Key Member & Role	Affiliation
Library Technology Services	Project Manager, Laura Morse	LTS
Information Technology Services	Technical Lead, TBD	LTS
Harvard Library	Project Lead, Lee LaFleur	HL
Harvard Library	Project Team, Reed Lowrie, Emily Bell, TBD	HL
Communications	Consultant, TBD	HL Communications

Cost and Estimated Schedule

Phase	Phase Start	Phase End	Completion Milestone
Planning & Vendor Coordination	Spring 2026	Summer 2026	Functional requirements delivered to vendors. Project plans created.
Development	Summer 2026	Winter 2027	Sprints for development complete.
Testing	Fall 2026	Winter 2027	Testing of agile releases complete.
Communication & Training	Winter 2027	Winter 2027	Community awareness campaign complete
Rollout out	Spring 2027	Summer 2027	System is live and available to all users.

Assumptions, Constraints, Dependencies, and Risks

Assumptions

Harvard Library Website and key Springshare data will be used for knowledge base, as determined by partner teams and vendors.
External partnership with other peers and vendors will inform tool.

Constraints

Resource availability across HL and LTS teams.
Coordination across multiple vendor platforms.

Dependencies

Creation of MCP and or API service by Springshare.
Development partnership with Clarivate on proof of concept for tool.

Risks

Delays in vendor completion of data processes.
Resourcing for Springshare data cleanup, if needed.