Driving Business Innovation: Latest Generative AI Advancements & Success Story

Are you ready to revolutionize how you handle data? Join us for a webinar where we’ll bring you up to speed with the latest advancements in Generative AI technology and discover how leveraging FME with tools from giants like Google Gemini, Amazon, and Microsoft OpenAI can supercharge your workflow efficiency.

During the hour, we’ll take you through:

Guest Speaker Segment with Hannah Barrington: Dive into the world of dynamic real estate marketing with Hannah, the Marketing Manager at Workspace Group. Hear firsthand how their team generates engaging descriptions for thousands of office units by integrating diverse data sources—from PDF floorplans to web pages—using FME transformers, like OpenAIVisionConnector and AnthropicVisionConnector. This use case will show you how GenAI can streamline content creation for marketing across the board.
Ollama Use Case: Learn how Scenario Specialist Dmitri Bagh has utilized Ollama within FME to input data, create custom models, and enhance security protocols. This segment will include demos to illustrate the full capabilities of FME in AI-driven processes.
Custom AI Models: Discover how to leverage FME to build personalized AI models using your data. Whether it’s populating a model with local data for added security or integrating public AI tools, find out how FME facilitates a versatile and secure approach to AI.
We’ll wrap up with a live Q&A session where you can engage with our experts on your specific use cases, and learn more about optimizing your data workflows with AI.

This webinar is ideal for professionals seeking to harness the power of AI within their data management systems while ensuring high levels of customization and security. Whether you're a novice or an expert, gain actionable insights and strategies to elevate your data processes. Join us to see how FME and AI can revolutionize how you work with data!

1. Driving Business Innovation:
Latest Generative AI
Advancements & Success Story

2. Chris
Berger

Customer Solutions
Team Lead
Safe Software
Dmitri
Bagh

Technical Specialist
Safe Software
Hannah
Barrington

Marketing Manager
Workspace Group
Oliver
Morris

Business Director
Tensing
Meet the Presenters

3. Welcome to Livestorm.
A few ways to engage with us during the webinar:


Audio issues? Click this for 4 simple
troubleshooting steps.

4. Agenda
1Introduction
2Customer Story: Workspace Group
3Locally Run LLM & Embedding
4Google Gemini Connector
5Resources & Next Steps
6Q&A
Agenda

5. 1
Introduction

6. Harness the transformative
power of AI.
Gain actionable insights for efficiency in
data integration strategy.



Introduction

7. Common Barriers to AI Adoption
●Technology Complexity: Difficulty grasping & applying advanced AI systems.
●Skill Shortages: Insufficient expertise to optimize AI use.
●Security Concerns: High stakes around data privacy and security.
●Resource Limits: Financial and time constraints hindering AI adoption.
●Change Resistance: Company hesitancy to adopt disruptive technologies.

Introduction

8. FME simplifies AI
integration, bridging
resource constraints with
less cost & complexity.











Introduction

9. One platform, two technologies
FME Form FME Flow
Data Movement and transformations
(“ETL”) workflows are built here.
Brings life to FME Form workflows
FME Flow Hosted
Safe Software managed FME Flow
fme.safe.com/platform
FME Enterprise Integration Platform
Safe & FME

10. 2
Customer
Success Story
Description generation using Generative AI

11. WORKSPACE
GROUP

12. Introductions
Workspace Group
-Leading provider of flexible office
spaces in London
-Own > 5m sq.ft. of space
-77 properties in London
-Home to 4,000+ diverse
businesses

13. Introductions

14. Hannah Barrington
Marketing Manager
- Property marketing/acquisition
-Building launches
-Focus buildings - boost occupancy
-Branded collateral
-Shop window
-Building signage

15. Some
Background

16. ●Upgraded DAM from Sharepoint to Bynder
-Sharepoint - Images in standard folder format
-Bynder - Images searchable using tags
- System integration
- SEO/keywords

●Lack of efficiency

●Manual process


Image migration
Background

17. ●SharePointOnlineConnector
●HttpCaller
○Importing the OpenAPI spec
●> 8000 images loaded so far


Image migration
Background
FME Form

18. ●Images synced to CRM,
website and marketing
outputs
●Sped up manual unit
photography process
Result
Background

19. ●Only had illegible architect plans & areas of buildings had changed
●Plans were in varying formats - PDFs, DWGs
●Various failed methods over 3+ years
Floor plan creation - Background
Background

20. ●PDF / CAD to PDF
●Over 800 units created in 1 month


Floor Plan creation
Background
FME Form

21. ●Producing whole buildings at a time
●18 buildings complete
●Remaining building rollout in
coming months

Result

22. The
Challenge

23. Slide Title
Generate unit
descriptions for
thousands of
Workspace office
units across over
70 buildings in
London

Goal Block Key
Unit descriptions using FME + AI
Result
Datasets spread
across DAM
catalogue, MS
Dynamics CRM,
PDFs, Google
Sheets and Web
pages

Using FME to
retrieve datasets
and put AI
enabled
transformers to
work

POC has
generated over a
hundred unit
descriptions, with
further integrations
planned

24. ●Currently short and generic default copy
●Written by copywriter as and when units need a boost
●No extra detail around;
-Style
-Views
-Flooring
-Windows/doors
-Amenities eg. AC
-Natural light
-Specifications
-Size

Unit Descriptions

25. Multi-Modal Data
●Floorplans in PNG
●Unit images
●Descriptions (Unit, Building, Local Area)
Unit descriptions using FME + AI

26. Data Sources Readers / Transformers
●Digital Asset Management
catalogue (DAM) - Bynder
●Dynamics CRM
●File system - PDFs
●Google Sheets
●Public Web Pages
●HttpCaller
●PDF Reader
●Google Sheets Reader
●PNG (Portable Network
Graphics) Writer
●OpenAIVisionConnector
●AnthropicVisionConnector
Unit descriptions using FME + AI

27. Demo

29. ●FME great at reading
disparate formats
●FME also is great at
orchestration
●Perfecting prompts is
crucial
●Use multiple models and
tools



Results
Unit descriptions using FME + AI

30. ●Unit photos – from 1 year to
2 weeks
●Floorplans – 3.5 years to
~4,000 units within next 6
months
●Unit descriptions – POC
done in 1 week, possible to
generate 4000 unit
descriptions in hours



Benefits
Unit descriptions using FME + AI

31. ●Write directly out to MS Dynamics CRM
●Use the building CADs to generate more
information - how far away are:
○Phone booths, Drinks points
○Meeting rooms/Break out spaces
●Building area descriptions

Next Steps
Unit descriptions using FME + AI

32. 3
Locally run LLM,
Embedding

33. Slide Title
Create metadata
for a photo
archive

Goal Block Key
Using AI for Image Metadata Creation
Result
●Manual work is too
laborious
●Photos may
include sensitive
contents
Using locally-run
LLM
The metadata
database and
visualization
tools were
created

34. What is Ollama?
Ollama is a tool that helps you
use AI models directly or in
your apps easily, without
needing to set up complicated
systems.

https://ollama.com/

35. Photo
Archive
Local Models for
Sensitive Data:
LLama3+Llava
ChatGPT for
non-sensitive
components:
HTML, CSS,
JavaScript
FME to rule
them all

36. ●Ask ChatGPT for metadata
database structure and SQL
statements for DB creation
●Read a folder of photos
●Ask LLava model for photo
descriptions
●Ask Llama3 model to
generate tags and creative
titles
●Normalize data and store in
database
●Ask ChatGPT for
HTML/CSS/JS
●Create a workspace to
transform DB to JSON
●Visualize your data
Photo Archive Creation
and Visualization

37. Photo Archive Creation
and Visualization
●Ask ChatGPT for metadata
database structure and SQL
statements for DB creation
●Read a folder of photos
●Ask LLava model for photo
descriptions
●Ask Llama3 model to
generate tags and creative
titles
●Normalize data and store in
database
●Ask ChatGPT for
HTML/CSS/JS
●Create a workspace to
transform DB to JSON
●Visualize your data

38. Photo Archive Creation
and Visualization
●Ask ChatGPT for metadata
database structure and SQL
statements for DB creation
●Read a folder of photos
●Ask LLava model for photo
descriptions
●Ask Llama3 model to
generate tags and creative
titles
●Normalize data and store in
database
●Ask ChatGPT for
HTML/CSS/JS
●Create a workspace to
transform DB to JSON
●Visualize your data

39. Visualize your data

View results

Can you see what’s wrong with the tags?
Photo Archive Creation
and Visualization

40. Lessons learned
●You can protect your data with locally run LLM
●Use the best tool available for generating
open/non-sensitive contents
●Prompt writing becomes an essential skill

Problems discovered
●Inconsistent prompt following (three requests
per image add up)
●Huge number of tags

Updated Solution
●1 request for all tags, then - assigning images to
tags, not tags to images

View results

Next problem
●What if we have 20,000 photos?

41. Embeddings
Python
Vector database
ChatGPT
Ollama
FME to rule
them all

42. ●Embeddings is a way to
turn words into
numbers.
●We can test how
embeddings work with a
collection of fictional
facts to feed into a
model.
●A small dataset is good
for understanding how
embeddings works and
shows whether the
results are relevant and
accurate.


You will need Python libraries:
Ollama, ChromaDB*
Ollama Embeddings Tutorial

Simple Embedding Tests. What is a Tigerat?

43. Embedding transformer descriptions

44. Embedding transformer descriptions

45. Demo

46. ●Extract object information
●Turn object info and questions into embeddings
●Ask questions
●Save and visualize answers
Embedding Revit data. Workspace

47. ●Extract Object ID from answers
●Join with geometries
●Send to Data Inspector
Embedding Revit data. Visualization

48. 4
Google
Gemini
Connector

49. Google Vertex Model Garden

50. Google Vertex AI
Authentication: OAuth

Integration: REST API

Vertex Multimodal support
●Language: Gemini
●Vision: Imagen
●Speech: Universal Speech Model (USM)
GoogleGeminiConnector

51. Demo

52. GoogleGeminiConnector Demo

53. OpenAIChatGPTConnector
●OpenAI: GPT-4o
●Azure OpenAI Model Catalog:
Easy access to 1,672 models
FME & Cloud Based AI Services
AmazonBedrockConnector
●AI21 labs
●Claude
●Cohere
●Stable Diffusion
●Amazon Titan
●Llama
GoogleGeminiConnector
●Gemini 1.5 Pro
●Model Garden: Easy access to 165
models




Discover more on the FME Hub!

54. FME & Cloud Based AI Services
Google Vertex Model Garden
●Deploy over 700,000 models from
HuggingFace with API access via
the Google Gemini OAuth web
connection & the HTTPCaller




Discover more on the FME Hub!

55. 6
Resources

56. Get our Ebook
Spatial Data for the
Enterprise

fme.ly/gzc


Guided learning
experiences at your
fingertips
academy.safe.com


FME Academy
Resources
Check out how-to’s &
demos in the knowledge
base
support.safe.com
Knowledge Base Webinars
Upcoming &
on-demand webinars

safe.com/webinars

57. Check out
our podcasts
on-demand.
featuring special guest
speakers over at EM360

Resources

58. 7
Next Steps

59. We’d love to help you get
started.
Get in touch with us at
[email protected]
Experience the
FME Accelerator
Contact Us
A world where data is not just a
commodity but a catalyst for
real change.
fme.safe.com/accelerator



Next Steps

60. Claim Your Community Badge &
Dive into the new Community!
●Get community badges for watching
webinars
●community.safe.com
●Today’s code: A7CKPP

Join the Community today!
Next Steps

61. 8
Q&A

62. Thank You
Recap of Next Steps

1Join the FME Community
2Contact us
3Experience the FME Accelerator

Please fill out our
webinar survey