AI-Powered Data Integration with FME
AI-Powered Data Integration with FME
Structured data extraction is a key capability of Large Language Models (LLMs). Providers like OpenAI and Gemini offer APIs that transform unstructured content into clean JSON outputs. But LLMs alone aren’t enough. Without robust tools to automate, validate, and scale the process, you’re left with fragile scripts and manual clean-up. That’s where FME comes in -seamlessly integrating generative AI into real-world data workflows.
Sign up for our FME Form (previously FME Desktop) focused AI training course and take your investment in FME into the exciting topic of orchestrating artificial intelligence! On our AI-Powered Data Integration with FME course, we help you navigate FME's ability to connect you to the AI service you need, sometimes utilising more than one service to get the job done.
- Use any AI provider, from OpenAI and Gemini to local models, in one workflow.
- Keep sensitive data in-house by running open-source AI models locally with Ollama.
- Turn PDFs, unknown schemas and messy records into clean, structured, usable data.
- Go beyond chatbots, combining computer vision, generative AI and embeddings in real workflows.
- Hands-on throughout, taught by certified FME trainers who build AI-powered data workflows for clients.
Goals
- Understand the generative AI landscape and how providers such as OpenAI and Google Gemini differ, so you can choose the right model for each task.
- Connect workspaces to AI services through their APIs, writing effective prompts and returning structured JSON that can be parsed reliably.
- Run open-source models locally with Ollama, so you can use AI on sensitive data without relying on external services.
- Clean, standardise and enrich messy data with AI, moving beyond manual review and rules-based logic.
- Handle unknown schemas and unstructured sources, using AI to map incoming data to your own structure and extract tables from PDFs.
- Orchestrate multi-model workflows, combining computer vision and generative AI, and using embeddings to search and match content by meaning rather than keyword.
This hands-on, one-day course from Avineon Tensing shows you how to build Generative AI into your FME workspaces, turning messy, unstructured and unpredictable content into clean, structured data. Building on foundational FME skills, you’ll start with how AI services are reached through their APIs. You’ll then connect to models from providers such as OpenAI and Google Gemini, and to open-source models running locally with Ollama for cases where security and data protection matter most. Along the way you’ll learn to write effective prompts and get back structured JSON you can rely on, then use that to clean and enrich real-world data. The course then moves on to harder problems. These include mapping unknown schemas to your own structure, extracting tables from PDFs, orchestrating Roboflow computer vision models alongside generative AI, and using embeddings to search content by meaning rather than keyword. You’ll leave with practical, multi-model workflows you can put to use straight away.
Who is this course for?
This course is intended for an existing user of FME Form, but don't worry, you only need to have scratched the surface since this course is a deep dive into AI as much as it is FME. If you're keen to understand how AI and ML can help you move faster and add-value to data workflows, it's for you. Typical candidates are:
- FME users eager to supercharge their automations with AI.
- Data management professionals wanting to tame their unstructured data and extract value from it.
- Those seeking scalable, cost-effective solutions to classify, extract and validate datasets using AI and Machine Learning (ML).
- Anyone wanting to keep up with this fast-paced topic and spend some time and immersion in AI automations!
TESTIMONIAL: "Learning new and valuable skills that I have been able to apply and benefit from immediately. Great course, thank you!" Mark Lloyd, Managing Director, Pear Technology Services Ltd
TESTIMONIAL: "The course was well organized and ran smoothly without any technical issues." Konrad Poplawski, Senior GIS Specialist, Rosen
Our Approach
To ensure a flexible training experience, we deliver this course in the following formats:
- 1 full day of focused training between 09:15 and 16:30.
- Bundled with content from other courses as part of a bespoke offering just for your organisation.
If you'd like something bespoke or would like to chat about the details, just get in touch by completing the form below and adding some extra detail to the 'please tell us more' text box.
Preknowledge
Attendees should have a basic working knowledge of FME Form, including understanding of readers, writers and transformers.
Attendees should understand how to:
- Know how to run an FME Workspace
- Work with readers and writers for common spatial data formats including Shapefile, GeoJSON and file geodatabase
- Apply basic transformers for data manipulation
This course does not require prior AI or machine learning experience.
Duration
1 day training course
Register now for an FME Training course
Training events:
- November 20th 2026 - FULL
- December 10th 2026
- March 2nd 2027
- May 12th 2027
- July 20th 2027
- October 13th 2027