Use Case: Automating SLS Community Gallery Finalisation and Review Preparation

 

Use Case: Automating SLS Community Gallery Finalisation and Review Preparation

Problem / Workflow Addressed

When SLS modules are prepared for submission to the Community Gallery, officers often need to perform a series of repetitive checks and refinements before publication. These activities can include:

• Reviewing content consistency
• Checking instructional text and metadata
• Ensuring resources are properly organised
• Verifying readiness before a module is shared with a wider audience
• Preparing modules for review and approval workflows

These tasks are important but time-consuming, especially when multiple modules need to be processed within a short period.

How ChatGPT Edu Was Used

Using ChatGPT Edu together with a custom Codex skill hosted in the codexSkill repository, an AI-assisted workflow was created to support the finalisation of SLS Community Gallery modules. The repository contains several SLS-focused skills designed to automate MOE SLS workflows, including module transfer, community review and related publishing activities. [github.com], [github.com]

The workflow leverages:

• Structured prompts and instructions embedded within the Codex skill
• ChatGPT Edu's reasoning capability to analyse module data
• Automated generation of recommendations and checks
• Consistent processing of module information according to predefined rules

Instead of manually inspecting every aspect of a module, officers can use AI to perform much of the preliminary review and finalisation work.

Value / Improvement Observed

Key benefits include:

1. Time Savings

Routine validation and review preparation activities can be completed significantly faster than a fully manual process.

2. Greater Consistency

AI follows the same checking procedure every time, reducing variation between reviewers.

3. Reduced Cognitive Load

Officers can focus on pedagogical quality and learning design instead of administrative checking.

4. Scalability

The same workflow can be applied to many modules without requiring additional manpower.

Examples / Evidence

Examples of practical outcomes include:

• Faster preparation of Community Gallery submissions.
• More consistent review recommendations across modules.
• Reduced effort in identifying issues before submission.
• Creation of reusable workflows that other divisions can adopt.

The broader codexSkill repository already demonstrates a growing collection of reusable SLS workflow automations, including Community Gallery review and module transfer processes, suggesting that the approach can be replicated across multiple operational tasks. [github.com]

Challenges Encountered

Limited Access to Structured Data

Some SLS information may not be easily accessible for automated processing.

Prompt and Workflow Maintenance

As SLS features evolve, prompts and skills require periodic updates.

User Confidence

Officers may need time to develop trust in AI-generated recommendations and learn how to review outputs critically.

Governance Considerations

AI-generated outputs still require human validation before publication or approval decisions.

Further Support Required

To scale this initiative, the following support would be valuable:

1. Expanded ChatGPT Edu access for HQ and school officers who manage SLS content.
2. A shared repository of approved MOE AI workflows so that divisions do not duplicate effort.
3. Official guidance on AI-assisted SLS workflows including governance and quality assurance practices.
4. Training on building reusable Codex skills so non-technical officers can automate their own workflows.
5. Integration opportunities with SLS APIs or structured exports to reduce manual handling.

Closing Statement

This proof of concept demonstrates how ChatGPT Edu can move beyond content generation into workflow automation. By codifying repeated Community Gallery finalisation tasks into reusable AI skills, officers can reduce administrative workload, improve consistency and focus more attention on pedagogy and learner outcomes. The experience also highlights the potential for a growing library of reusable MOE workflows that can be shared and adapted across divisions. [github.com], [github.com]

Potential ask to the group: We would welcome opportunities to collaborate with other divisions exploring similar AI workflow automation use cases, particularly around reusable skill development, governance guidance, and integration with existing MOE platforms.