Generative AI

Slide 1Claris Explorer Lab

Generative AI Reports in FileMaker

Slide 2About Matt

• Passionate FileMaker trainer
• Expert FileMaker developer
• Certified in FM versions 7 to present
• Founded 3 companies
• FBA Awards, conference sessions, Meetups
• ClarisTalk AI podcast with Cris Ippolite

Slide 3Navarre Training

• www.navarre.training
• matt@navarre.training
• Schedule a free zoom with me

Slide 4FileMaker Champion Series

• Live, interactive, skills-based, current best practices
• AI Integration Class starts October 20
• Learn best practices
• New features including AI integration
• 10 hours of class time + 1:1 instruction with each student

Slide 5Agenda

• A few slides
• Interactive training ~ 35 minutes
• Q&A

Slide 6Goals

• Definitions
• Embedding vs. Generative AI
• Discuss Examples
• Interactive Exercises

Slide 7Definitions

• LLM (Large Language Model) – e.g. GPT-4, GPT-5; the AI engines behind generative text
• API Key – Authenticate and call AI services
• Token – a chunk of text “Cost per million tokens”
• Context Window – how many tokens the model can handle at once.

Slide 8Definitions

• Prompts – instructions you give to the model
• System prompts – role/behavior
• User prompts – tasks/questions
• Embeddings & Vectors
• Numeric representations of text meaning
• Used for semantic search and similarity

Slide 9What is Generative AI?

• Creates content based on:
• Pre-trained data – vast, but can hallucinate
• Your data + pre-trained data
• Low hallucination rate
• High quality / meaningful result
• FileMaker Pro for the win

Slide 10Generative AI vs. Embedding

• Embedding (separate free class)
• Condense bits of meaning from text or image
• Store embedded data in FileMaker
• Semantic Search on embedded data
• Runs locally - no API call
• Same results every time

Slide 11Generative AI vs. Embedding

• Generative AI
• Calls a model (ie gpt-5-nano) for each use
• Creates text, image, code, html, etc.
• Analyze, summarize, expand, translate, etc.
• Different results every time

Slide 12Think of it like…

• 1,000 interns…
• With doctoral degrees
• In every subject
• In every language
• …but they are interns
• And require careful guidance

Slide 13Examples

• Summarize long text
• Translate to any language
• Write a product description (given specs)
• Draft personalized emails
• Explain technical info in plain language
• Create a bar chart or report

Slide 14Let’s Explore

• Start from provided sample data file
• Configure AI Account
• System Prompt
• User Prompt
• Generate Response from Model
• Perform Find by Natural Language

Slide 15Exercises

• Hello World
• Translate Text
• Find by Natural Language
• Create a chart
• Create an HTML report

Slide 16To the interactive class…