Plan du cours
Introduction
-
What is business automation with ChatGPT?
-
Models, agents, tools, and workflows
-
Fixed workflows vs agentic processes
-
Choosing the right level of autonomy
-
The role of human review and decision-making
-
Beyond a single ChatGPT conversation: reusable workflows and recurring execution
Understanding Context and Memory
-
Chat context vs persistent memory
-
Explicitly stored process state
-
Preserving decisions and important information between runs
-
Context limitations and loss of earlier information
-
Continuing work across multiple conversations
-
Distinguishing permanent rules from current process status
Organizing Work in ChatGPT
-
The risks of combining too many tasks and changing requirements in one thread
-
Recognizing missing decisions and inconsistent outputs
-
Splitting work into smaller, clearly defined tasks
-
Passing goals, sources, decisions, and results between tasks
-
When to use separate conversations or workflows
Working with Plugins, Projects, and Spaces
-
What plugins, projects, and Spaces are used for
-
Organizing reusable instructions and tools
-
Connecting workflows to business information sources
-
Organizing related conversations and shared materials
-
Sharing documentation and results
-
Preparing the working environment for a business process
Building a Workflow with Memory
-
Defining the workflow objective
-
Identifying inputs, steps, outputs, and acceptance criteria
-
Recording process state between executions
-
Using previous decisions in subsequent runs
-
Handling errors and incomplete execution
-
Resuming a workflow after interruption
-
Building a prototype workflow for a selected business task
Scheduling and Repeatable Execution
-
Manual vs recurring workflow execution
-
Defining execution frequency and time zone
-
Notification and stopping rules
-
Handling cases where no new data is available
-
Re-running a workflow with saved state
-
Comparing repeated executions
Evaluating Workflow Quality and Repeatability
-
Defining quality criteria
-
Checking required fields, numbers, sources, and output format
-
Identifying duplicate or inconsistent results
-
Distinguishing acceptable wording differences from process errors
-
Understanding sources of variability in AI-generated results
-
Testing the workflow on representative business cases
Practical Workshop
-
Selecting a recurring business task
-
Designing the workflow
-
Defining process memory and state
-
Running and testing the workflow
-
Comparing multiple executions
-
Identifying errors and improvement opportunities
-
Preparing the automation for practical use
Troubleshooting
-
Missing or incomplete context
-
Incorrect or outdated process state
-
Inconsistent results between executions
-
Missing data or unavailable sources
-
Duplicate processing
-
Failed or partially completed workflow runs
-
Tool or account feature limitations
Summary and Next Steps
-
Reviewing the completed workflow prototype
-
Identifying tasks suitable for automation
-
Defining quality and acceptance criteria
-
Planning deployment in daily work
-
Identifying opportunities for a more advanced second-day version with code-supported processing and agent development
Pré requis
-
What delegates should know prior to the course
- Basic familiarity with ChatGPT
- No programming knowledge is specified as a prerequisite.
- Delegates should have a computer, internet connection, and a ChatGPT account with access to the features used during the workshops.
- Availability of plugins, Spaces, and scheduling/automation features should be checked before the course, as access can depend on account and organizational settings.
Audience
The course is addressed to:
- Specialists and managers using ChatGPT in their daily work
- Process owners and business analysts responsible for improving team workflows
- People preparing reports, summaries, documents, data compilations, and recurring business updates
- Team leaders implementing AI and organizing work with shared information sources
A short catalogue-style version could be:
Prerequisites: Basic knowledge of ChatGPT. No programming experience required.
Audience: Managers, specialists, process owners, business analysts, reporting/documentation professionals, and AI implementation leaders.
Nos clients témoignent (1)
Capable de s'adapter en fonction des suggestions du public, par exemple en créant un scénario d'agent IA en temps réel.
Brett McLaren - Zoll Itamar
Formation - ChatGPT for Productivity: A Beginner’s Guide
Traduction automatique
