Slide 1

Slide 1 text

Real-World AI Automation with Lambda & Bedrock

Slide 2

Slide 2 text

๐ŸŒŽ What Does Real-World Mean?

Slide 3

Slide 3 text

https://www.coatue.com/blog/company- update/coatues-2023-emw-conference

Slide 4

Slide 4 text

๐Ÿ”ต Expand Product Suite with AI

Slide 5

Slide 5 text

๐Ÿ”ต Expand Product Suite with AI GenAI Sprinkles

Slide 6

Slide 6 text

๐ŸŸข Pivot Entire Business

Slide 7

Slide 7 text

๐ŸŸข Pivot Entire Business https:/ /www.databricks.com/dataaisummit

Slide 8

Slide 8 text

๐ŸŸฃ Start Fresh New Venture

Slide 9

Slide 9 text

๐Ÿ”ด New Tailwinds for Existing

Slide 10

Slide 10 text

๐Ÿ”ด New Tailwinds for Existing

Slide 11

Slide 11 text

๐Ÿ”ด New Tailwinds for Existing https:/ /dev.to/aws-heroes/rags-to-riches-part-1-generative-ai-retrieval-4pd7

Slide 12

Slide 12 text

Q: Where Should Your Company Start? https:/ /hbr.org/2023/09/where-should-your-company-start-with-genai

Slide 13

Slide 13 text

A: Use The WINS Framework https:/ /hbr.org/2023/09/where-should-your-company-start-with-genai (W)ords (I)mages (N)umbers (S)ounds More re๏ฌned & precise category of knowledge work for GenAI adoption. Lever: Enhances productivity in non-digital ๏ฌelds Crucible: Essential for tech-heavy sectors Balcony: Min. impact on manual labor sectors. Next: Improves engagement in traditional Industries

Slide 14

Slide 14 text

โ— ๐Ÿ—บ New Tailwinds for your Existing Business โ— ๐ŸŽ› WINS Holding a Lever: โ€œSellingโ€ part of SG&A โ— โญ WINS Next in Line: Digital Front Door or Engagement ๐ŸŒŽ Real-World AI Summary ๐Ÿง 

Slide 15

Slide 15 text

๐Ÿค– AI Agents & Automation

Slide 16

Slide 16 text

โ— Digital assistants similar to Siri or Alexa but more advanced. โ— Uses Generative Pre-Trained (GPT) models to Reason & Act. โ— Can use Data and Tools to help inform decision making. โ— Can automate routine or large complex business tasks. โ— Fast to market and easy to augment with new data. โ— No upfront investments in compute training or data science. โ— Can ๏ฌne-tune pre-trained models to increase accuracy. ๐Ÿฆพ What are AI Agents? ๐Ÿง 

Slide 17

Slide 17 text

I AM HERE! ๐Ÿฆพ What are AI Agents? ๐Ÿง  https:/ /blog.langchain.dev/openais-bet-on-a-cognitive-architecture/

Slide 18

Slide 18 text

๐ŸฆฟOpenAIโ€™s Custom GPTs https:/ /openai.com/blog/introducing-gpts

Slide 19

Slide 19 text

๐ŸฆฟOpenAIโ€™s Custom GPTs https:/ /openai.com/blog/introducing-gpts

Slide 20

Slide 20 text

๐Ÿฆพ What are AI Agents? ๐Ÿง 

Slide 21

Slide 21 text

๐Ÿฆพ What are AI Agents? ๐Ÿง 

Slide 22

Slide 22 text

Real-World AI Agent Example ๐Ÿค–

Slide 23

Slide 23 text

๐ŸŽˆ Real-World AI Automation

Slide 24

Slide 24 text

๐Ÿ” Real-World AI Automation

Slide 25

Slide 25 text

๐Ÿ” Real-World AI Automation

Slide 26

Slide 26 text

๐Ÿ” Real-World AI Automation

Slide 27

Slide 27 text

โž• Bedrock is Simple. Easy API Usage. Serverless Billing. โž• Anthropic Claude Models are Amazing! Bedrock PE Guide. ๐šซ Wrapped API Access lacks new Tools & System Prompts. ๐šซ Token Usage is in CloudWatch Metrics vs. API Responses. ๐šซ Moving from Token to Provisioned has HUGE $$$ Gaps. ? Value or Not in Bedrockโ€™s Agents & Knowledge Products? ? New Local Development Patterns. Dataset-Driven. ๐Ÿชจ Real-World Bedrock Learnings https:/ /docs.google.com/presentation/d/1tjvAebcEyR8la3EmVwvjC7PHR8gfSrcsGKfTPAaManw/edit?usp=sharing

Slide 28

Slide 28 text

No content