Overview of ChatGPT: A GenAI Solution for Schools, Academies, and Multi-Academy Trusts
In November 2024, the team undertook an experiment with GenAI by creating a simple ChatGPT, powered by generative AI (GenAI) technology using publically available content on GOV.UK and Department of Education public websites.
Core Knowledge Base using publically available content
DfE frameworks online https://www.gov.uk/guidance/find-a-dfe-approved-framework-for-your-school
- Procurement guidance onlinehttp://GOV.UK
- Public Contracts Regulations 2015 online https://www.legislation.gov.uk/uksi/2015/102/contents
- DfE framework documents https://www.gov.uk/government/publications/department-for-education-framework-documents
- Guidance on Get Help Buying for Schools https://www.gov.uk/guidance/get-help-buying-for-schools
- Buying for schools: how to buy what you need https://www.gov.uk/guidance/buying-procedures-and-procurement-law-for-schools
GenAI represents a transformative tool in education procurement. Specifically tailored for schools, academies, and multi-academy trusts (MATs), this iteration—known as the "School Buying Professional" GPT—offers specialized expertise to support procurement professionals. It simplifies complex purchasing tasks, fosters compliance with Department for Education (DfE) guidance, and promotes value for money in safeguarding public funds.
This document explores the features of ChatGPT, its capabilities for supporting school procurement professionals, and the GenAI technology behind its development. Additionally, it examines how integrating custom knowledge bases and Department for Education content enhances its utility.
Generative AI (GenAI) and GPT Technology
What is GenAI?
Generative AI refers to systems capable of generating human-like text, images, code, or other content using advanced machine learning algorithms. Models like OpenAI’s GPT (Generative Pre-trained Transformer) are built to predict and generate coherent, contextually accurate content based on input data.
Key Features of GPT Models
- Natural Language Understanding (NLU): GPT can process, analyze, and generate human-like text in a conversational tone.
- Pre-training and Fine-tuning: Models are pre-trained on vast datasets and fine-tuned for specific use cases, such as education procurement.
- Context Awareness: GPT excels in understanding the context and delivering nuanced answers.
- Customizability: The technology can be tailored to meet the unique requirements of organizations, such as adding DfE frameworks or procurement policies.
Features of the School Buying Professional GPT
The School Buying Professional GPT provides specialized tools and resources to assist school procurement professionals.
Core Features:
Expert Procurement Guidance
- Step-by-step advice on procurement law, drafting tenders, and engaging suppliers.
- Tailored responses to questions about frameworks, suppliers, and purchasing strategies.
Simplified Processes
- Automates document creation, such as templates for tender specifications and contracts.
- Breaks down complex tasks into actionable steps.
Approved Framework Navigation
- Recommends DfE-approved frameworks like Crescent Purchasing Consortium (CPC), YPO, and ESPO.
- Directs users to pre-approved suppliers that ensure compliance and value for money.
Content Integration
- Uses Department for Education content to provide official guidance.
- Adheres to http://GOV.UK style standards for clarity and accuracy.
Knowledge Base Integration
- Can be enhanced by integrating custom knowledge bases containing local policies or past procurement data.
- Tailors responses to specific school or MAT contexts.
Training and Compliance Tools
- Helps procurement professionals understand compliance requirements.
- Offers links to training resources for continuous professional development.
How ChatGPT Supports School Procurement Professionals
For Schools
- Small-Scale Procurement: Supports smaller budgets by suggesting efficient, low-cost frameworks for one-off purchases.
- Streamlined Compliance: Ensures purchases align with DfE requirements, saving administrative time.
For Academies
- Centralized Resources: Provides academies with centralized advice on managing procurement across multiple departments.
- Framework Alignment: Guides academies in leveraging approved frameworks for recurring expenses like cleaning services, IT, or catering.
For Multi-Academy Trusts (MATs)
- Strategic Procurement: Helps MATs pool resources and negotiate better deals across their schools.
- Tailored Policies: Integrates trust-specific procurement strategies into its advice.
- Collaboration Tools: Promotes collaboration by creating standardized tender templates and specifications.
Enhancing Procurement with Department for Education Content
The integration of DfE content ensures that ChatGPT delivers accurate, compliant, and actionable advice tailored to UK schools. Key applications of DfE content:
- Procurement Law Guidance: Provides up-to-date interpretations of the Public Contracts Regulations 2015.
- Framework Recommendations: Links users to DfE-approved frameworks, ensuring cost-effective and compliant purchasing.
- Sustainability Focus: Encourages the use of sustainable suppliers, particularly for catering and facilities management.
- Case Studies and Insights: Uses real-world examples from the Buying for Schools Blog to illustrate best practices.
Compliance with GOV UK Standards The use of DfE content ensures adherence to the GOV UK style guide, prioritizing clarity, simplicity, and user-centric advice.
Custom Knowledge Base Integration
By incorporating custom knowledge bases, the GPT’s functionality can be significantly enhanced:
- Localized Advice: Integrate school or trust-specific policies for tailored responses.
- Historical Data Usage: Provide insights based on previous procurement activities.
- Enhanced Accuracy: Address unique challenges faced by individual organizations.
Use Case: Integrating a School’s Procurement History
For a school with a history of catering procurement issues, a custom knowledge base could:
- Track supplier performance.
- Highlight successful frameworks previously used.
- Recommend alternative suppliers based on past challenges.
Examples of How ChatGPT Supports Procurement Tasks
Drafting Tender Specifications
ChatGPT simplifies tender creation by asking guiding questions:
- What is your budget?
- What outcomes do you expect?
- Have you considered approved frameworks?
It then generates tailored tender templates based on these answers.
Finding Approved Suppliers
Using tools like the Find a DfE-approved Framework, ChatGPT connects schools to pre-vetted suppliers.
Creating Procurement Timelines
Outlines clear timelines for procurement processes, ensuring adherence to legal requirements.
Guiding Framework Selection
Recommends suitable frameworks for specific services, such as catering (YPO) or IT procurement (CPC).
Benefits of Using ChatGPT for Schools and MATs
Cost and Time Savings
- Reduces administrative overhead by automating repetitive tasks.
- Streamlines supplier selection and tender drafting.
Improved Compliance
- Ensures adherence to legal requirements and DfE guidance.
Enhanced Decision-Making
- Offers data-driven insights and recommendations.
Accessible Expertise
- Makes complex procurement advice available to all school professionals.
Future Potential: Enhanced Functionality and Collaboration
Integration with Other Systems
- Combine ChatGPT with financial management software for real-time budget analysis.
- Integrate with project management tools to track procurement timelines.
Community Building
- Facilitate collaboration between schools and MATs to share best practices.
Expanded Content Sources
- Include insights from local councils and sustainability organizations to enrich advice.
Next Steps
ChatGPTs and similar GenAI technologies, could an invaluable tool for school procurement professionals, offering tailored support to navigate complex purchasing processes. By leveraging GenAI technology and Department for Education content, it simplifies compliance, promotes cost-efficiency, and enhances decision-making. The integration of custom knowledge bases further strengthens its ability to provide localized, context-specific advice.
For schools, academies, and MATs, ChatGPT is not just a technological innovation; it is a strategic partner in delivering high-quality education procurement.
Next Steps for Building a Finessed GenAI GPT for Schools and MAT Procurement
Developing a more refined GenAI GPT tailored for school procurement professionals requires a systematic approach to enhance its capabilities, integrate additional data, and rigorously test its effectiveness. Below is a roadmap for advancing the tool, focusing on iterative improvement, user feedback, and accuracy assessment.
1. Define the Enhanced Scope and Objectives
Goals for the Improved GPT
- Provide tailored, localized advice by integrating diverse knowledge bases.
- Increase accuracy and reliability in procurement guidance.
- Test the tool's utility, benefits, and user experience with real-world applications.
- Establish the GPT as a trusted partner for procurement decisions.
Expanded Use Cases
- Scenario-Specific Guidance: Handle nuanced procurement challenges, such as sustainability criteria, urgent procurements, and high-value contracts.
- Collaborative Features: Enable collaboration among MATs and schools to share resources and templates.
- Advanced Analytics: Provide data-driven insights into past purchases and supplier performance.
2. Enrich the Knowledge Base
Integrate Comprehensive Data Sources To enhance its depth, the GPT must incorporate a broader set of structured and unstructured data, including:
- Department for Education (DfE) Resources: Policies, procurement frameworks, and compliance guidelines.
- Case Studies: Real-world procurement examples from schools and MATs.
- Local Policies: Regional and school-specific procurement rules.
- Supplier Performance Data: Metrics on supplier reliability and satisfaction from approved frameworks like CPC, YPO, and ESPO.
Example of Knowledge Base Architecture
Core Knowledge Base
- DfE frameworks
- http://GOV.UK procurement guidance
- Public Contracts Regulations 2015
Supplemental Knowledge Base
- School and MAT procurement history
- Local council and regional policies
- Insights from sustainability organizations
User-Generated Content
- Feedback from procurement professionals
- Templates and documents shared by users
3. Develop a Testing Framework
Key Testing Areas
- Accuracy: Validate GPT’s advice against DfE standards and procurement laws.
- Relevance: Assess if responses meet specific user needs.
- Ease of Use: Measure user satisfaction with the interface and response clarity.
- Compliance: Ensure guidance adheres to legal and ethical standards.
User Testing Phases
Alpha Testing
- Limited deployment with a small group of procurement professionals.
- Test basic functionalities and gather preliminary feedback.
Beta Testing
- Broader deployment across schools, academies, and MATs.
- Include diverse user groups, from small schools to large trusts.
- Test advanced features, such as data-driven insights and custom knowledge integration.
Pilot Programs
- Partner with select MATs to implement the GPT in live procurement scenarios.
- Collect feedback on decision-making, time savings, and compliance.
Metrics for Evaluation
- User Engagement: Frequency of use and task completion rates.
- Time Efficiency: Reduction in time spent on procurement tasks.
- Accuracy: Percentage of advice aligned with procurement best practices.
- Cost Savings: Financial benefits achieved through improved procurement strategies.
4. Enhance the GPT’s Features
Personalization and Customization
- User Profiles: Allow users to set preferences (e.g., budget limits, preferred frameworks).
- Contextual Insights: Use historical data to inform future procurement decisions.
Integration Capabilities
- ERP Systems: Sync with financial and project management systems for holistic procurement oversight.
- Real-Time Updates: Include live updates from DfE and framework providers.
Advanced Functionality
- Scenario Analysis: Offer simulations of procurement outcomes based on different strategies.
- Compliance Checker: Validate draft tenders or contracts against procurement law.
5. Implement Feedback Loops
User Feedback Collection
- Surveys and interviews with procurement professionals during testing phases.
- Built-in feedback tools in the GPT interface (e.g., “Rate this Response” buttons).
Iterative Refinement
- Use feedback to address gaps, such as missing data sources or misunderstood user needs.
- Continuously update the GPT’s knowledge base to reflect changes in regulations or frameworks.
6. Measure Benefits and Accuracy
Assess Benefits
- Efficiency Gains: Track time saved by procurement professionals using the GPT.
- Financial Impact: Measure cost reductions achieved through optimized procurement.
- User Satisfaction: Evaluate whether the GPT meets the expectations of its primary users.
Accuracy Validation
- Compare GPT-generated advice with expert reviews.
- Monitor errors or misinterpretations and refine algorithms accordingly.
7. Explore the Long-Term Role of the GPT
Strategic Integration into Schools and MATs
- Establish the GPT as a central resource for procurement training.
- Position the tool as a repository of best practices and shared resources across MATs.
Continuous Improvement
- Plan for regular updates to incorporate evolving regulations, technology, and user needs.
- Build partnerships with framework providers to ensure direct integration of updates.
Work for 2025
The path to a more refined GenAI GPT lies in an iterative approach, leveraging user feedback and comprehensive data integration. By aligning with user needs and legal requirements, the GPT can become an indispensable tool for schools, academies, and MATs.
Immediate Next Steps:
- Develop the enriched knowledge base with diverse data inputs.
- Begin alpha testing with a focus group of procurement professionals.
- Implement feedback loops for continuous iteration and improvement.
By following this roadmap, the GPT will evolve into a robust, reliable, and indispensable asset for the education sector, ensuring compliance, efficiency, and value for money in procurement practices.