SlideShare a Scribd company logo
Welcome to ServerlessToronto.org
1
Introduce Yourself:
- Where from? Why are you here?
- Looking for work or Offering work?
Help us serve you better: bit.ly/slsto
“AI for Lean Startup” with Pankaj Upreti, Lagna360
18 - 18:10 Networking, Opening remarks
18:10 ~ 19:00 Langa360.com AI Journey
19:00 - 19:20 Q&A
19:20 - 19:30 Manning Publications raffle
What is Serverless Toronto about?
2
Serverless became New Agile & Mindset
#1 We started as Back-
end FaaS Developers
who enjoyed 'gluing
together' other people's
APIs and Managed
Services
#3 We're obsessed
with creating business
value (meaningful
Products), focusing on
Outcomes/Impact –
NOT Outputs
#2 We build bridges
between Serverless
Community (“Dev leg”),
and Front-end, Voice-First
& UX folks (“UX leg”)
#4 Achieve agility NOT by
“sprinting” faster but working
smarter (by using bigger
building blocks & less Ops)
1
2
3
4
Serverless is a State of Mind…
3
Way too often, we – the IT folks,
have obsession with “pimping up
our cars” (infrastructure / code /
pipelines) instead of “driving
business” forward & taking them
places ☺
... It is a way to focus on business value.
The Serverless Mindset
4
It can be applied to any Tech stack, even On-Prem
Jared Short:
1. If the platform has it, use it
2. If the market has it, buy it
3. If you can reconsider requirements, do it
4. If you have to build it, own it.
Ben Kehoe: Serverless is about how you make
decisions, not about your choices.
Upcoming ServerlessToronto.org Meetups
5
Monday, August 12 at 5pm EST Fall and Winter talks?
http://youtube.serverlesstoronto.org/
Knowledge Sponsor
1. Select *any* e-Book, Video course, or liveProject you want!
2. Add it to your shopping cart (no more than 1 item in the cart)
3. Raffle winners will send me the emails (used in Manning portal),
4. So the publisher can move it to your Dashboard – as if purchased.
5. 45%-off code for us: TORONTO24, go to https://mng.bz/DdZV to use it
Fill out the Survey to win: bit.ly/slsto
7
Feature Presentation:
AI:YOUR STARTUP
SIDEKICK
LEVERAGING AI TO BOOTSTRAP A LEAN STARTUP
AGENDA
What we are covering
How to make AI your trusty sidekick in
bootstrapping a startup. Use cases,
examples from the Journey of lagna360,
using AI to operate lean
What We’re Skipping
Deep AI algorithms, code walk-through’s,
step by step guilds
2
INTRODUCTION
My Background
• Pankaj Upreti, Founder of Lagna360
• Enterprise architect / technologist – worked in
software industry for 25+ years. Large enterprise
to startups. Project management to coding, pre-
sales to architecture. FORTRAN to Virtual Reality
• Passionate about astrology, archeology and
ancient history.
Introduction to Lagna360
• AI-driven astrology platform
• Personalized horoscope readings, NASA precision
• Global userbase, exponential growth.
3
WHY AI FOR STARTUPS?
1. Importance for Cash-Strapped Startups
• Startups need to stretch every dollar.
• AI can help automate tasks and reduce costs.
• Fast iteration is crucial for startup success
2. Ride the Wave
• New models emerging every few weeks.
• AI is the hot topic in tech circles everywhere.
3. Promises of AI
• AI promises intelligence and efficiency.
• But does it deliver on these promises?
• Let's find out how AI can truly help your startup.
4
FOCUS ON CONSUMER
FACING SAAS
Not a lemonade stand
• AI is less useful in physical business
Why SAAS?
• Build a startup that can run from
anywhere (even your couch!).
• Flexible, scalable, and good for
leveraging AI.
Examples of Virtual Startups
• Personalized health advice apps
• Social platforms, like Facebook, X
5
COGNITIVE LOAD?
A Small Brain Attached to Your Own Brain
• AI acts as an extra appendage to offload tiresome
parts of thinking.
• With this "extra brain," you can reduce mental
fatigue and work more efficiently.
A Tool, Not a Replacement
• Using AI doesn't mean you can play golf while AI
does all the work.
• AI is like a power tool or forklift; it helps reduce
load but doesn't replace the mechanic.
• With reduced cognitive load, you can work more
hours with less fatigue and increase productivity.
6
THE JOURNEY OF LAGNA360
The challenges faced by a typical SAAS startup
7
INITIAL CONCEPT
Inspiration
Personal interest in astrology and AI
The Seed
Created the core engine using 6GB of NASA to make
accurate astronomy calculations, backbone of astrology
Eureka Moment
Realizing the potential to combine AI with Vedic astrology
to create a unique, high-accuracy service.
8
PRODUCT CHALLENGE
1. Analyzing Astrological Texts
• Several dozen books on astrology
• Extracting and tuning prediction algorithms
2. Analyzing Historical Charts
• Over 150,000 charts to study
• Comparing predictions with actual life events
of well-known individuals
3. Developing the Core Prediction Engine
• Integrating insights from texts and charts
• Ensuring high accuracy and personalization
in predictions
9
ARCHITECTURE
CHALLENGE
Scalability and Performance
• Handle a growing number of users
• Optimizing performance
Continuous Improvement
• Integrating new AI models and tools
• Maintaining compatibility across updates
Data Security and Privacy
• Protecting sensitive user data
• Ensuring compliance with data protection regulations
Cross-Platform Compatibility
• Works seamlessly across devices
• Consistent user experience across all channels
10
MARKETING CHALLENGE
• Content Creation
• Graphic Design
• Video Production
• SEO Optimization
• Social Media Management
• Advertising Campaigns
• Analytics and Reporting
• Customer Support
11
AI - INTHE REAL WORLD
Highlight the various AI tools and techniques used to stay lean.
12
CODING ASSISTANCE
Tools
• CodeWhisperer
• Github Co-pilot
• chatGPT
Impact
• 25% speed improvement (3 instead of 4 devs)
• Mental energy conservation (12 hrs instead of 8)
• Repetitive boring tasks, refactoring, experimenting
• Overall reduce developer requirement by 50%
13
AUTOMATEDTESTING
Tools Used
• ChatGPT
• Github co-pilot
• DeepCodeAI – static code analysis
Impact
• Early bug detection and prevention
• Improved code quality and reliability
• Reduced manual testing effort
• Reduced mental load (less fatigue)
• 50% less time spent with open code
14
INTEGRATIONTESTING
Tools Used
• Selenium
• chatGPT used to help write selenium tests, inject
realistic user data
Impact
• Simplified complex testing
• Reduced testing time by 40%
• Enhanced integration reliability
• Improved test coverage and accuracy
• Less time with open code, reduced fatigue
15
MARKETING CONTENT
Tools Used
• Anthropic Claud
• ChatGPT
• Custom agentic flow
Impact
• Increased content production speed
• Enhanced creativity and engagement
• Consistent and high-quality content
• Eliminate need to hire content writer
16
IMAGE GEN –VISUAL
CONTENT
Tools Used
• DALL-E
• Canva with AI features
• Photoshop with AI features
Impact
• Rapid creation of high-quality images
• Customizable visuals for various marketing needs
• Enhanced visual appeal and engagement
• Reduced reliance on graphic designers
• Time otherwise spent on coding diverted to using these
tools, eliminate need to hire designers.
17
GA4 ANALYTICS: INSIGHTS
WITH AI
Tools Used
• ChatGPT (file upload + screen capture)
• There are newer tools available, still
reviewing
Impact
• Enhanced understanding of user behavior
• Data-driven decision making
• Improved website performance
• Identification of key growth opportunities
18
19
Database analytics
In god we trust – for everything else we
need data
• https://aws.amazon.com/blogs/database
/how-to-perform-advanced-analytics-
and-build-visualizations-of-your-amazon-
dynamodb-data-by-using-amazon-athena
• Once data is in Athena, use ChatGPT to
help compose queries
• Prepare the GPT context, with
knowledge of your business and
objectives.. Then ask for queries.
SOCIAL MEDIA AUTOPILOT
Tools Used
• Pictory video generator
• Hootsuite
Impact
• Not my core skillset, managing without social media agent
by using automation and AI
• More content, good quality content, constant and with
cadence.
20
LARGE CONTEXT AND
HALLUCINATION
CHALLENGE
• Example
- In chat window user could ask
- Insurance policy related question
- Mortgage policy related question
- Assume both categories have knowledge base
of 100,000 tokens each like a Harry Potter
novel
- Assume answer is not just text, but a detailed
calculation is required as part of the result
- How do you return result with no
hallucination?
21
22
Why? Cost Management
• Handling very large contexts that require deep analysis.
• More cost-effective than using more expensive models directly.
• Reduces hallucinations and maintains performance.
Pattern Overview
• Step 1: AWS Lambda -> Classification Model
• Initial call to determine user intent.
• Step 2: Context Preparation
• Lambda fetches data from session/backend to prepare a compressed,
targeted context.
• Step 3: AWS Lambda -> LLM Call
• Make LLM call with the targeted context.
Advantages
• Like RAG but without the need for a vector database.
• More suitable for dynamic data processing, calculations, and DB retrievals.
• Cost-effective and efficient in extracting performance from lower-cost models.
CLASSIFICATION ENHANCED PROCEDURAL GENERATION
VIDEO CONTENT &
ADVERTISING
Tools Used
• Pictory, HeyGen, D-ID, RunWay
• Capcut , RVC libraries
Impact
• Increased content production speed
• Enhanced creativity and engagement
• Consistent and high-quality content
• Eliminate need to hire video creator
23
DEMO BREAK
9
9
AI MAKESTHIS EASIERTHANYOUTHINK
2022 – only 2 years ago, it would require outsourcing to make a video of this quality
Tools to make this commercial
•Google analytics data on user geography, age , gender, interests and usage patterns
•ChatGPT to analyze GA report and brainstorm targeting ideas
•ChatGPT prepares the concept / script
•Dall-e for consistent image generation
•Runway ML – image to video generation
•Pictory – takes the script and generates video with voice over and scene transitions
•Capcut for final editing.
Sound like too much work? Remember has been your sidekick till now, leads to reduced “mental load”,
leaves the human mind free to create and innovate without fatigue.
KEYTAKE AWAYS
3 Things to Remember
26
1. CONTINUOUS
IMPROVEMENT
- Always be improving and
adapting. The landscape of
technology and business is ever-
evolving, and so must our
strategies and approaches.
- Embrace change, seek out
learning opportunities, and
continuously strive to enhance
your skills and workflows.
27
2. KNOWLEDGE GAPS & AI
- recognize what you do not know.
AI is a powerful tool, but it cannot
replace the insights and
understandings that you have not
yet acquired.
- AI cannot help in tasks that you
cannot “somewhat” do yourself
28
3. LESS EFFORT MORE HOURS
Productivity vs. Hours: Technology does not
always reduce the number of hours worked;
instead, it makes those hours more
productive and focused.
Quality of Work: Reduced cognitive strain
enables higher quality work, as individuals
can spend more time on complex problems
without increased stress.
29
THANKYOU!
@LAGNA_360
HTTPS://WWW.LAGNA360.COM
www.ServerlessToronto.org
Reducing the gap between IT and Business needs

More Related Content

PDF
Accelerate ML Deployment with H2O Driverless AI on AWS
PDF
Power to the People: A Stack to Empower Every User to Make Data-Driven Decisions
PPTX
Maintainable Machine Learning Products
PDF
Digital Transformation - Why you need to embrace it now
PDF
Women in Automation - Intro to Studio Session 1
PDF
Building a Data Driven Company
PPTX
SPSNYC2019 - What is Common Data Model and how to use it?
PDF
Market_Cloud_AI_Capabilities_POC Demo.pdf
Accelerate ML Deployment with H2O Driverless AI on AWS
Power to the People: A Stack to Empower Every User to Make Data-Driven Decisions
Maintainable Machine Learning Products
Digital Transformation - Why you need to embrace it now
Women in Automation - Intro to Studio Session 1
Building a Data Driven Company
SPSNYC2019 - What is Common Data Model and how to use it?
Market_Cloud_AI_Capabilities_POC Demo.pdf

Similar to AI - Your Startup Sidekick (Leveraging AI to Bootstrap a Lean Startup).pdf (20)

PDF
🌺 Women in Automation Series: Intro to Studio ▶ Session 1
PPTX
Machine Learning with GraphLab Create
PPTX
Group 3 slide presentation
PDF
8base Hyperledger Miami Meetup 20180719
PDF
8base Hyperledger Miami Meetup Presentation
PPTX
Supercharge_AI_Presentation-fornbrgginrt .pptx
PPTX
It Consulting & Services - Black Basil Technologies
PDF
Application Modernization
PPTX
Serverless Toronto helps Startups
PPTX
Getting Started With Dato - August 2015
PPTX
Best Practices for API Adoption - WIP Factory presentation for AnyPresence we...
PDF
How Celtra Optimizes its Advertising Platform with Databricks
PDF
Bridging the Gap: Analyzing Data in and Below the Cloud
PPTX
Magnolia 6 release walkthrough
PPTX
Dashlane Mission Teams
PPTX
Making sense of microservices, service mesh, and serverless
PDF
Data Visualization and the Art of Self-Reliance
PPTX
Content Analytics Studio – The visualization, machine learning and applicatio...
PPTX
Data science tools of the trade
PDF
Aw (3) webinar serverless-fisher-rymer
🌺 Women in Automation Series: Intro to Studio ▶ Session 1
Machine Learning with GraphLab Create
Group 3 slide presentation
8base Hyperledger Miami Meetup 20180719
8base Hyperledger Miami Meetup Presentation
Supercharge_AI_Presentation-fornbrgginrt .pptx
It Consulting & Services - Black Basil Technologies
Application Modernization
Serverless Toronto helps Startups
Getting Started With Dato - August 2015
Best Practices for API Adoption - WIP Factory presentation for AnyPresence we...
How Celtra Optimizes its Advertising Platform with Databricks
Bridging the Gap: Analyzing Data in and Below the Cloud
Magnolia 6 release walkthrough
Dashlane Mission Teams
Making sense of microservices, service mesh, and serverless
Data Visualization and the Art of Self-Reliance
Content Analytics Studio – The visualization, machine learning and applicatio...
Data science tools of the trade
Aw (3) webinar serverless-fisher-rymer
Ad

More from Daniel Zivkovic (20)

PDF
'The Art & Science of LLM Reliability - Building Trustworthy AI Systems' by M...
PDF
All in AI: LLM Landscape & RAG in 2024 with Mark Ryan (Google) & Jerry Liu (L...
PDF
Canadian Experts Discuss Modern Data Stacks and Cloud Computing for 5 Years o...
PDF
Opinionated re:Invent recap with AWS Heroes & Builders
PDF
Google Cloud Next '22 Recap: Serverless & Data edition
PDF
Conversational Document Processing AI with Rui Costa
PDF
How to build unified Batch & Streaming Pipelines with Apache Beam and Dataflow
PDF
Gojko's 5 rules for super responsive Serverless applications
PDF
Retail Analytics and BI with Looker, BigQuery, GCP & Leigha Jarett
PDF
What's new in Serverless at AWS?
PDF
Intro to Vertex AI, unified MLOps platform for Data Scientists & ML Engineers
PDF
Empowering Developers to be Healthcare Heroes
PDF
Get started with Dialogflow & Contact Center AI on Google Cloud
PDF
Building a Data Cloud to enable Analytics & AI-Driven Innovation - Lak Lakshm...
PDF
Smart Cities of Italy: Integrating the Cyber World with the IoT
PDF
Running Business Analytics for a Serverless Insurance Company - Joe Emison & ...
PDF
This is my Architecture to prevent Cloud Bill Shock
PDF
Lunch & Learn BigQuery & Firebase from other Google Cloud customers
PDF
Azure for AWS & GCP Pros: Which Azure services to use?
PDF
Serverless Evolution during 3 years of Serverless Toronto
'The Art & Science of LLM Reliability - Building Trustworthy AI Systems' by M...
All in AI: LLM Landscape & RAG in 2024 with Mark Ryan (Google) & Jerry Liu (L...
Canadian Experts Discuss Modern Data Stacks and Cloud Computing for 5 Years o...
Opinionated re:Invent recap with AWS Heroes & Builders
Google Cloud Next '22 Recap: Serverless & Data edition
Conversational Document Processing AI with Rui Costa
How to build unified Batch & Streaming Pipelines with Apache Beam and Dataflow
Gojko's 5 rules for super responsive Serverless applications
Retail Analytics and BI with Looker, BigQuery, GCP & Leigha Jarett
What's new in Serverless at AWS?
Intro to Vertex AI, unified MLOps platform for Data Scientists & ML Engineers
Empowering Developers to be Healthcare Heroes
Get started with Dialogflow & Contact Center AI on Google Cloud
Building a Data Cloud to enable Analytics & AI-Driven Innovation - Lak Lakshm...
Smart Cities of Italy: Integrating the Cyber World with the IoT
Running Business Analytics for a Serverless Insurance Company - Joe Emison & ...
This is my Architecture to prevent Cloud Bill Shock
Lunch & Learn BigQuery & Firebase from other Google Cloud customers
Azure for AWS & GCP Pros: Which Azure services to use?
Serverless Evolution during 3 years of Serverless Toronto
Ad

Recently uploaded (20)

PPTX
Transform Your Business with a Software ERP System
PPTX
Agentic AI Use Case- Contract Lifecycle Management (CLM).pptx
PPTX
ISO 45001 Occupational Health and Safety Management System
PDF
Adobe Illustrator 28.6 Crack My Vision of Vector Design
PDF
Nekopoi APK 2025 free lastest update
PDF
Design an Analysis of Algorithms II-SECS-1021-03
PPTX
ai tools demonstartion for schools and inter college
PDF
top salesforce developer skills in 2025.pdf
PDF
Digital Strategies for Manufacturing Companies
PPT
Introduction Database Management System for Course Database
PPTX
Online Work Permit System for Fast Permit Processing
PPTX
Operating system designcfffgfgggggggvggggggggg
PPTX
history of c programming in notes for students .pptx
PPTX
Odoo POS Development Services by CandidRoot Solutions
PPTX
ManageIQ - Sprint 268 Review - Slide Deck
PDF
Raksha Bandhan Grocery Pricing Trends in India 2025.pdf
PDF
Softaken Excel to vCard Converter Software.pdf
PDF
How to Choose the Right IT Partner for Your Business in Malaysia
PPTX
Agentic AI : A Practical Guide. Undersating, Implementing and Scaling Autono...
PDF
Addressing The Cult of Project Management Tools-Why Disconnected Work is Hold...
Transform Your Business with a Software ERP System
Agentic AI Use Case- Contract Lifecycle Management (CLM).pptx
ISO 45001 Occupational Health and Safety Management System
Adobe Illustrator 28.6 Crack My Vision of Vector Design
Nekopoi APK 2025 free lastest update
Design an Analysis of Algorithms II-SECS-1021-03
ai tools demonstartion for schools and inter college
top salesforce developer skills in 2025.pdf
Digital Strategies for Manufacturing Companies
Introduction Database Management System for Course Database
Online Work Permit System for Fast Permit Processing
Operating system designcfffgfgggggggvggggggggg
history of c programming in notes for students .pptx
Odoo POS Development Services by CandidRoot Solutions
ManageIQ - Sprint 268 Review - Slide Deck
Raksha Bandhan Grocery Pricing Trends in India 2025.pdf
Softaken Excel to vCard Converter Software.pdf
How to Choose the Right IT Partner for Your Business in Malaysia
Agentic AI : A Practical Guide. Undersating, Implementing and Scaling Autono...
Addressing The Cult of Project Management Tools-Why Disconnected Work is Hold...

AI - Your Startup Sidekick (Leveraging AI to Bootstrap a Lean Startup).pdf

  • 1. Welcome to ServerlessToronto.org 1 Introduce Yourself: - Where from? Why are you here? - Looking for work or Offering work? Help us serve you better: bit.ly/slsto “AI for Lean Startup” with Pankaj Upreti, Lagna360 18 - 18:10 Networking, Opening remarks 18:10 ~ 19:00 Langa360.com AI Journey 19:00 - 19:20 Q&A 19:20 - 19:30 Manning Publications raffle
  • 2. What is Serverless Toronto about? 2 Serverless became New Agile & Mindset #1 We started as Back- end FaaS Developers who enjoyed 'gluing together' other people's APIs and Managed Services #3 We're obsessed with creating business value (meaningful Products), focusing on Outcomes/Impact – NOT Outputs #2 We build bridges between Serverless Community (“Dev leg”), and Front-end, Voice-First & UX folks (“UX leg”) #4 Achieve agility NOT by “sprinting” faster but working smarter (by using bigger building blocks & less Ops) 1 2 3 4
  • 3. Serverless is a State of Mind… 3 Way too often, we – the IT folks, have obsession with “pimping up our cars” (infrastructure / code / pipelines) instead of “driving business” forward & taking them places ☺ ... It is a way to focus on business value.
  • 4. The Serverless Mindset 4 It can be applied to any Tech stack, even On-Prem Jared Short: 1. If the platform has it, use it 2. If the market has it, buy it 3. If you can reconsider requirements, do it 4. If you have to build it, own it. Ben Kehoe: Serverless is about how you make decisions, not about your choices.
  • 5. Upcoming ServerlessToronto.org Meetups 5 Monday, August 12 at 5pm EST Fall and Winter talks? http://youtube.serverlesstoronto.org/
  • 6. Knowledge Sponsor 1. Select *any* e-Book, Video course, or liveProject you want! 2. Add it to your shopping cart (no more than 1 item in the cart) 3. Raffle winners will send me the emails (used in Manning portal), 4. So the publisher can move it to your Dashboard – as if purchased. 5. 45%-off code for us: TORONTO24, go to https://mng.bz/DdZV to use it Fill out the Survey to win: bit.ly/slsto
  • 8. AI:YOUR STARTUP SIDEKICK LEVERAGING AI TO BOOTSTRAP A LEAN STARTUP
  • 9. AGENDA What we are covering How to make AI your trusty sidekick in bootstrapping a startup. Use cases, examples from the Journey of lagna360, using AI to operate lean What We’re Skipping Deep AI algorithms, code walk-through’s, step by step guilds 2
  • 10. INTRODUCTION My Background • Pankaj Upreti, Founder of Lagna360 • Enterprise architect / technologist – worked in software industry for 25+ years. Large enterprise to startups. Project management to coding, pre- sales to architecture. FORTRAN to Virtual Reality • Passionate about astrology, archeology and ancient history. Introduction to Lagna360 • AI-driven astrology platform • Personalized horoscope readings, NASA precision • Global userbase, exponential growth. 3
  • 11. WHY AI FOR STARTUPS? 1. Importance for Cash-Strapped Startups • Startups need to stretch every dollar. • AI can help automate tasks and reduce costs. • Fast iteration is crucial for startup success 2. Ride the Wave • New models emerging every few weeks. • AI is the hot topic in tech circles everywhere. 3. Promises of AI • AI promises intelligence and efficiency. • But does it deliver on these promises? • Let's find out how AI can truly help your startup. 4
  • 12. FOCUS ON CONSUMER FACING SAAS Not a lemonade stand • AI is less useful in physical business Why SAAS? • Build a startup that can run from anywhere (even your couch!). • Flexible, scalable, and good for leveraging AI. Examples of Virtual Startups • Personalized health advice apps • Social platforms, like Facebook, X 5
  • 13. COGNITIVE LOAD? A Small Brain Attached to Your Own Brain • AI acts as an extra appendage to offload tiresome parts of thinking. • With this "extra brain," you can reduce mental fatigue and work more efficiently. A Tool, Not a Replacement • Using AI doesn't mean you can play golf while AI does all the work. • AI is like a power tool or forklift; it helps reduce load but doesn't replace the mechanic. • With reduced cognitive load, you can work more hours with less fatigue and increase productivity. 6
  • 14. THE JOURNEY OF LAGNA360 The challenges faced by a typical SAAS startup 7
  • 15. INITIAL CONCEPT Inspiration Personal interest in astrology and AI The Seed Created the core engine using 6GB of NASA to make accurate astronomy calculations, backbone of astrology Eureka Moment Realizing the potential to combine AI with Vedic astrology to create a unique, high-accuracy service. 8
  • 16. PRODUCT CHALLENGE 1. Analyzing Astrological Texts • Several dozen books on astrology • Extracting and tuning prediction algorithms 2. Analyzing Historical Charts • Over 150,000 charts to study • Comparing predictions with actual life events of well-known individuals 3. Developing the Core Prediction Engine • Integrating insights from texts and charts • Ensuring high accuracy and personalization in predictions 9
  • 17. ARCHITECTURE CHALLENGE Scalability and Performance • Handle a growing number of users • Optimizing performance Continuous Improvement • Integrating new AI models and tools • Maintaining compatibility across updates Data Security and Privacy • Protecting sensitive user data • Ensuring compliance with data protection regulations Cross-Platform Compatibility • Works seamlessly across devices • Consistent user experience across all channels 10
  • 18. MARKETING CHALLENGE • Content Creation • Graphic Design • Video Production • SEO Optimization • Social Media Management • Advertising Campaigns • Analytics and Reporting • Customer Support 11
  • 19. AI - INTHE REAL WORLD Highlight the various AI tools and techniques used to stay lean. 12
  • 20. CODING ASSISTANCE Tools • CodeWhisperer • Github Co-pilot • chatGPT Impact • 25% speed improvement (3 instead of 4 devs) • Mental energy conservation (12 hrs instead of 8) • Repetitive boring tasks, refactoring, experimenting • Overall reduce developer requirement by 50% 13
  • 21. AUTOMATEDTESTING Tools Used • ChatGPT • Github co-pilot • DeepCodeAI – static code analysis Impact • Early bug detection and prevention • Improved code quality and reliability • Reduced manual testing effort • Reduced mental load (less fatigue) • 50% less time spent with open code 14
  • 22. INTEGRATIONTESTING Tools Used • Selenium • chatGPT used to help write selenium tests, inject realistic user data Impact • Simplified complex testing • Reduced testing time by 40% • Enhanced integration reliability • Improved test coverage and accuracy • Less time with open code, reduced fatigue 15
  • 23. MARKETING CONTENT Tools Used • Anthropic Claud • ChatGPT • Custom agentic flow Impact • Increased content production speed • Enhanced creativity and engagement • Consistent and high-quality content • Eliminate need to hire content writer 16
  • 24. IMAGE GEN –VISUAL CONTENT Tools Used • DALL-E • Canva with AI features • Photoshop with AI features Impact • Rapid creation of high-quality images • Customizable visuals for various marketing needs • Enhanced visual appeal and engagement • Reduced reliance on graphic designers • Time otherwise spent on coding diverted to using these tools, eliminate need to hire designers. 17
  • 25. GA4 ANALYTICS: INSIGHTS WITH AI Tools Used • ChatGPT (file upload + screen capture) • There are newer tools available, still reviewing Impact • Enhanced understanding of user behavior • Data-driven decision making • Improved website performance • Identification of key growth opportunities 18
  • 26. 19 Database analytics In god we trust – for everything else we need data • https://aws.amazon.com/blogs/database /how-to-perform-advanced-analytics- and-build-visualizations-of-your-amazon- dynamodb-data-by-using-amazon-athena • Once data is in Athena, use ChatGPT to help compose queries • Prepare the GPT context, with knowledge of your business and objectives.. Then ask for queries.
  • 27. SOCIAL MEDIA AUTOPILOT Tools Used • Pictory video generator • Hootsuite Impact • Not my core skillset, managing without social media agent by using automation and AI • More content, good quality content, constant and with cadence. 20
  • 28. LARGE CONTEXT AND HALLUCINATION CHALLENGE • Example - In chat window user could ask - Insurance policy related question - Mortgage policy related question - Assume both categories have knowledge base of 100,000 tokens each like a Harry Potter novel - Assume answer is not just text, but a detailed calculation is required as part of the result - How do you return result with no hallucination? 21
  • 29. 22 Why? Cost Management • Handling very large contexts that require deep analysis. • More cost-effective than using more expensive models directly. • Reduces hallucinations and maintains performance. Pattern Overview • Step 1: AWS Lambda -> Classification Model • Initial call to determine user intent. • Step 2: Context Preparation • Lambda fetches data from session/backend to prepare a compressed, targeted context. • Step 3: AWS Lambda -> LLM Call • Make LLM call with the targeted context. Advantages • Like RAG but without the need for a vector database. • More suitable for dynamic data processing, calculations, and DB retrievals. • Cost-effective and efficient in extracting performance from lower-cost models. CLASSIFICATION ENHANCED PROCEDURAL GENERATION
  • 30. VIDEO CONTENT & ADVERTISING Tools Used • Pictory, HeyGen, D-ID, RunWay • Capcut , RVC libraries Impact • Increased content production speed • Enhanced creativity and engagement • Consistent and high-quality content • Eliminate need to hire video creator 23
  • 32. 9 AI MAKESTHIS EASIERTHANYOUTHINK 2022 – only 2 years ago, it would require outsourcing to make a video of this quality Tools to make this commercial •Google analytics data on user geography, age , gender, interests and usage patterns •ChatGPT to analyze GA report and brainstorm targeting ideas •ChatGPT prepares the concept / script •Dall-e for consistent image generation •Runway ML – image to video generation •Pictory – takes the script and generates video with voice over and scene transitions •Capcut for final editing. Sound like too much work? Remember has been your sidekick till now, leads to reduced “mental load”, leaves the human mind free to create and innovate without fatigue.
  • 33. KEYTAKE AWAYS 3 Things to Remember 26
  • 34. 1. CONTINUOUS IMPROVEMENT - Always be improving and adapting. The landscape of technology and business is ever- evolving, and so must our strategies and approaches. - Embrace change, seek out learning opportunities, and continuously strive to enhance your skills and workflows. 27
  • 35. 2. KNOWLEDGE GAPS & AI - recognize what you do not know. AI is a powerful tool, but it cannot replace the insights and understandings that you have not yet acquired. - AI cannot help in tasks that you cannot “somewhat” do yourself 28
  • 36. 3. LESS EFFORT MORE HOURS Productivity vs. Hours: Technology does not always reduce the number of hours worked; instead, it makes those hours more productive and focused. Quality of Work: Reduced cognitive strain enables higher quality work, as individuals can spend more time on complex problems without increased stress. 29
  • 38. www.ServerlessToronto.org Reducing the gap between IT and Business needs