AI TRAINING
Creating Custom GPTs for Your Team
Design, deploy, and govern custom GPTs that make your team's workflows faster and more consistent.
What it covers
Participants learn to build purpose-specific GPTs inside ChatGPT Enterprise or Team—covering system instructions, knowledge file curation, external actions via APIs, and safe sharing patterns within an organisation. The programme combines hands-on GPT Builder sessions with governance best practices so every custom GPT is both useful and compliant. By the end, each participant ships at least one working GPT tailored to a real team workflow. Format is a focused 2-day workshop with pre-work and a post-session review call.
What you'll be able to do
- Write a production-ready system prompt with persona, constraints, and tone guidelines for a specific team use case
- Upload, chunk, and test knowledge files so a custom GPT retrieves accurate answers from internal documents
- Configure at least one external Action using an OpenAPI schema to pull live data into a GPT conversation
- Publish a custom GPT to a ChatGPT Team or Enterprise workspace with appropriate access controls
- Apply a repeatable review checklist to audit GPT outputs for accuracy, bias, and data-privacy compliance
Topics covered
- Anatomy of a custom GPT: instructions, persona, and capabilities
- Writing effective system prompts and behavioural guardrails
- Uploading and structuring knowledge files for retrieval accuracy
- Connecting external data sources via Actions and OpenAPI schemas
- Testing, iterating, and evaluating GPT output quality
- Sharing models within ChatGPT Team or Enterprise workspaces
- Access control, versioning, and deprecation policies
- Governance: data privacy, acceptable-use policies, and GDPR considerations
Delivery
Delivered in person or via live virtual sessions (Zoom/Teams) across two days of 6–8 hours each. Day 1 focuses on instruction design and knowledge files; Day 2 covers Actions, sharing, and governance. Approximately 70% of time is hands-on in GPT Builder. A pre-work pack (1–2 hours) asks participants to document one target workflow before arrival. A 90-minute follow-up review call is held 2–3 weeks post-workshop to troubleshoot live deployments. Each participant needs a ChatGPT Team or Enterprise seat; facilitator provides prompt templates and a governance checklist as take-home assets.
What makes it work
- Anchoring each GPT to a single, well-defined workflow rather than trying to build an all-purpose assistant
- Establishing an internal owner (a 'GPT steward') responsible for maintaining instructions and knowledge files over time
- Running a structured pilot with a small team before org-wide rollout to catch edge cases early
- Creating a simple internal catalogue of approved custom GPTs so teams can discover and reuse existing ones
Common mistakes
- Dumping large, unstructured documents into knowledge files and expecting accurate retrieval without chunking or testing
- Writing vague system instructions that produce inconsistent GPT behaviour across different users
- Sharing GPTs org-wide before running even basic output-quality checks, causing trust issues when errors surface
- Ignoring OpenAI's data-usage policies and inadvertently feeding sensitive customer data into a non-Enterprise plan
When NOT to take this
If your organisation has not yet secured a ChatGPT Team or Enterprise subscription and has no near-term plan to do so, this workshop delivers no immediately deployable output — prioritise foundational AI tool selection and procurement first.
Providers to consider
Sources
This training is part of a Data & AI catalog built for leaders serious about execution. Take the free diagnostic to see which trainings your team needs.