Course Outline
Introduction to AI Builder and Low-Code AI
- Overview of AI Builder features and typical use cases.
- Considerations regarding licensing, governance, and tenant-level configurations.
- Introduction to Power Platform integrations, including Power Apps, Power Automate, and Dataverse.
OCR and Form Processing: Handling Structured and Unstructured Documents
- Distinguishing between structured templates and free-form documents.
- Preparing training data: field labeling, ensuring sample diversity, and adhering to quality standards.
- Constructing an AI Builder form processing model and assessing extraction accuracy.
- Post-processing extracted data through validation, normalization, and error management.
- Lab exercise: Performing OCR extraction from mixed form types and integrating it into a processing workflow.
Prediction Models: Classification and Regression
- Defining problem scopes: qualitative (classification) versus quantitative (regression) objectives.
- Feature engineering and managing missing data within Power Platform workflows.
- Training, testing, and interpreting model performance metrics such as accuracy, precision, recall, and RMSE.
- Addressing model explainability and fairness in business contexts.
- Lab exercise: Developing a custom prediction model for churn scoring or numeric forecasting.
Integration with Power Apps and Power Automate
- Incorporating AI Builder models into canvas and model-driven apps.
- Developing automated flows to process extracted data and initiate business actions.
- Establishing design patterns for scalable and maintainable AI-driven applications.
- Lab exercise: Executing an end-to-end scenario involving document upload, OCR processing, prediction, and workflow automation.
Supplementary Process Mining Concepts (Optional)
- Utilizing Process Mining to discover, analyze, and improve processes via event logs.
- Leveraging Process Mining outputs to refine model features and automate improvement cycles.
- Case study: Integrating Process Mining insights with AI Builder to minimize manual exceptions.
Production Readiness, Governance, and Monitoring
- Data governance, privacy, and compliance protocols when processing sensitive documents with AI Builder.
- Managing the model lifecycle: retraining, version control, and performance tracking.
- Operationalizing models through alerts, dashboards, and human-in-the-loop validation.
Summary and Future Steps
Requirements
- Practical experience with Power Apps, Power Automate, or Power Platform administration.
- A solid understanding of data concepts, fundamental ML principles, and model evaluation methods.
- Proficiency in handling datasets, Excel/CSV exports, and performing basic data cleansing.
Target Audience
- Power Platform developers and solution architects.
- Data analysts and process owners aiming to leverage AI for automation.
- Business automation leaders concentrating on document processing and prediction scenarios.
Testimonials (3)
Practical and hands on labs on report developmemt using Power BI The labs were excellent and the trainer offered very good hands on sessions
Sinzala Sichaanji - Bank of Zambia
Course - Mastering Power Platform: Power Apps, Power Automate, DataVerse, Power BI, and Power Virtual Agents
We did quite complex examples, so we could get a feeling of how the real work with Power Automate Desktop can look like in the real world scenario.
Michal Strnad - MicroNova AG
Course - Microsoft Flow/Power Automate
Dynamic, adaptive, and informative
