Generative AI at SAP

Unit 2: Introduction to Generative AI

Introduction to generative AI
Objectives of Unit 2

In Unit 2, you will learn

▪ What’s behind all the generative AI news and its
impact on the economy
▪ Definitions of foundation models, large language
models, and generative AI
▪ Why generative AI is a new approach to artificial
intelligence
▪ Examples of generative AI

© 2023 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC 2

Introduction to generative AI
Generative AI is all over the news

ChatGPT Is a Tipping Point for AI

Source: Harvard Business Review, 14. December 2022

Congress to hold new AI hearings

as it works to craft safeguards
Source: Reuters, 8. September 2023

‘Video games are in for quite a trip’: How

generative AI could radically reshape gaming
Source: CNN, 23. October 2023

The Coming AI Economic Revolution

Source: Foreign Affairs, November/December 2023

How generative AI could add trillions to SAP prepares to add Joule

the global economy generative AI copilot across its apps
Source: World Economic Forum, 14. July 2023 Source: CIO, 26. September 2023

© 2023 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC 3

Introduction to generative AI
Generative AI will profoundly impact the economy and increase productivity

Generative AI could add

$2.6 to $4.4 trillion

incremental value annually to
the global economy Already widespread adoption of generative AI

of companies globally are using

33% generative AI regularly in at
least one business function

of organizations will increase their

$11-$17tn 40% investment in AI overall because
of advances in generative AI
productivity value from existing
generation of narrow AI by 2040
of organizations adopting AI have
60% seen some revenue increase

Source: McKinsey, State of AI in 2023, 1,684 participants globally

© 2023 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC
Source: McKinsey, The economic potential of generative AI: 4
The next productivity frontier (June 2023)
Introduction to generative AI
Definitions

Foundation models are neural networks

Foundation trained on large volumes of data using self-
LLM learning objective:
Models supervised learning that can be applied to
Predict the next word in text.
many tasks.

Large England
London is in _______
Large language models (LLMs)
Language
are a subcategory of foundation Two hours south of Toronto, London is
Models (LLMs)
models for text incl. computer code. Ontario
in _______

Generative AI can create novel output in … trained on billions of words, images,

OpenAI’s text, images, sound, or video based on lines of code, etc.
ChatGPT simple user input (called “prompts”).

ChatGPT is an application built on a large

language model created by the vendor
OpenAI. The model has been optimized for
dialog.
Generative AI
Multi-modal models can process prompts and
generate output in various formats, including text,
images, video, and speech.

© 2023 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC 5

Introduction to generative AI
Capabilities of foundation models

Example
Capabilities
Characteristics
Data ▪ Huge models, often measured in billions of
Classification,
parameters and trained on billions/trillions of tokens
entity extraction
▪ Stateless: respond only to data sent to the model,
Text but do not “remember” or “learn” from previous
interactions
Summarization
▪ Performance and capabilities, in general, scale as
Image models increase in size

TRAINING Search and Q&A

New approach to AI
Speech
▪ From supervised to self-supervised learning
Foundation Content creation
models ▪ From predefined to emerging capabilities
(images, music,
… ▪ From single-purpose to multi-scenario AI video…)
▪ From classifying AI to generative AI
Users can easily interact with the model without Code generation
requiring prior technical knowledge.

© 2023 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC 6

Introduction to generative AI
Foundation models come in all sizes and generally scale with more data and parameters

Trainable Parameters (in billions): Example LLMs by parameter size

1,200

175 1,000

800

Attention-based transformer
600
as neural network
architecture (2017)
400

200
1.6
0.1
0
GPT (2018) GPT-2 XL GPT-3 (2020) Mistral Llama Falcon Llama GPT-3 BLOOM Lum- PaLM Mega- GPT-
(2019) 7B 2 13B 40B 2 70B inous tron 4?

Foundation models scale with more compute, more data, and more parameters (model size)
→ Higher accuracy, adaptability, and generative performance.

But also a trend towards smaller models benchmarking on certain tasks as well as much larger models due
to better training data and methods.
© 2023 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC 7
Introduction to generative AI
Examples of generative AI

Generative AI processes prompts Stable Diffusion 2 (text-to-image multimodal model)

Prompt: explicit instructions or cues to guide
“Hasso Plattner, the founder of SAP,
generative AI models in producing desired
giving a high five to a robot.”
outputs Prompt

Output:

© 2023 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC 8

Introduction to generative AI
Examples of generative AI

Large Language Models: create text. Large Language Models: summarize text.

Write an email introducing our logistics Summarize key benefits from the following
company, Treasure Island Transports Ltd, to 1000-word document in less than 100
a retail prospect. Do it the style of a pirate. Prompt words. Use bullet points.
Prompt

Ahoy there, ▪ SAP Business AI offers transformative intelligence to

businesses through ready-to-use AI capabilities in all
We be Treasure Island Transports Ltd, masters of the seven
business processes such as lead-to-cash, design-to-
seas! Our fleet be as swift as the wind, and our crew as
operate, recruit-to-retire, and source-to-pay.
reliable as the North Star. We be ready to set sail with your
goods, ensuring they reach port safe and sound. So, hoist ▪ SAP Business AI allows businesses to extend AI
the anchor and join us for a voyage of smooth logistics and capabilities to meet their specific needs using the SAP
swift deliveries. Business Technology Platform.
Yarrrs sincerely, ▪ SAP Business AI provides trust and reliability through AI
capabilities built on stringent ethics policy and data privacy
[Your Name]
standards that enable responsible use of AI with full
Treasure Island Transports Ltd. transparency and compliance.

© 2023 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC 9

Introduction to generative AI
Examples of generative AI

Large Language Models: classify and extract text. Write working code based on a text description

Classify the sentence as negative, neutral,

or positive sentiment and extract the object.
Prompt “I loved your course on openSAP!”

Positive sentiment
The object being referred to in the sentence is “your course
on openSAP”.
Prompt
Code

© 2023 SAP SE or an SAP affiliate company. All rights reserved. ǀ PUBLIC 10

Thank you.
Contact information:

[email protected]
Follow all of SAP

www.sap.com/contactsap

© 2023 SAP SE or an SAP affiliate company. All rights reserved.

No part of this publication may be reproduced or transmitted in any form or for any purpose without the express permission of
SAP SE or an SAP affiliate company.
The information contained herein may be changed without prior notice. Some software products marketed by SAP SE and its
distributors contain proprietary software components of other software vendors. National product specifications may vary.
These materials are provided by SAP SE or an SAP affiliate company for informational purposes only, without representation or
warranty of any kind, and SAP or its affiliated companies shall not be liable for errors or omissions with respect to the materials.
The only warranties for SAP or SAP affiliate company products and services are those that are set forth in the express warranty
statements accompanying such products and services, if any. Nothing herein should be construed as constituting an additional
warranty.
In particular, SAP SE or its affiliated companies have no obligation to pursue any course of business outlined in this document or
any related presentation, or to develop or release any functionality mentioned therein. This document, or any related presentation,
and SAP SE’s or its affiliated companies’ strategy and possible future developments, products, and/or platforms, directions, and
functionality are all subject to change and may be changed by SAP SE or its affiliated companies at any time for any reason
without notice. The information in this document is not a commitment, promise, or legal obligation to deliver any material, code, or
functionality. All forward-looking statements are subject to various risks and uncertainties that could cause actual results to differ
materially from expectations. Readers are cautioned not to place undue reliance on these forward-looking statements, and they
should not be relied upon in making purchasing decisions.
SAP and other SAP products and services mentioned herein as well as their respective logos are trademarks or registered
trademarks of SAP SE (or an SAP affiliate company) in Germany and other countries. All other product and service names
mentioned are the trademarks of their respective companies.
See www.sap.com/trademark for additional trademark information and notices.