Workshop 3:
AI Integration Masterclass: Mapping Tools, Data, and Workflows
Week 3: Preparing your environment for successful AI implementation

by Haojun See

On The Ground - AI Adoption for Businesses and Individuals
Why Mapping Matters
Know Your Environment
Can't integrate AI without understanding existing tools and formats
Data Chaos = Poor Results
Disorganized inputs lead to ineffective AI outputs
Readiness Assessment
Evaluate how prepared your workspace is for AI integration
Data Types: Structured vs Unstructured
Structured Data
Spreadsheets, forms, tables
Easy for AI to parse
Clearly defined format
Unstructured Data
PDFs, freeform notes, emails
Needs summarizing or extraction
Requires more prompt engineering
Tool Stack and Handoffs
Gmail
Email communication
Word/PDF
Document creation
Zoom
Virtual meetings
Google Calendar
Scheduling
Most workflows fail at handoffs: too much retyping or manual formatting
This Week's Goal
Audit
Review your tool and data ecosystem
Identify
Find key points for AI intervention
Prepare
Select one workflow for integration
What Makes a Tool AI-Ready?
Text Exportability
Can you easily copy or export text?
Consistent Format
Does it use predictable structures?
Text-Based Outputs
Are final products primarily text?
Repeatable Tasks
Do you perform similar actions often?
Data Transformation Examples
Guided Audit Activity
Tool Inventory
List the 5 tools you use daily
Data Assessment
Identify what types of data live in each tool
Structure Analysis
Determine if data is structured or unstructured
Output Evaluation
Can this tool produce reusable text?
Document Challenges & Workarounds
Common Problems
  • PDFs with mixed layouts
  • Inconsistent formatting
  • Headers/footers interfering
Prompt Strategies
  • Process one paragraph at a time
  • Clean then summarize
Use specific pre-prompts to guide the AI assistant.
For example, you could start with a prompt like "Ignore headers and footers."
Focusing on the main content is crucial when working with an AI assistant. By directing the AI to ignore headers, footers, and other peripheral elements, you can ensure that it concentrates solely on the core information you need.
This kind of pre-prompt helps the AI understand the specific part of the content you want it to work on. It sets clear boundaries and expectations, allowing the AI to hone in on the most relevant and important aspects of the material.
Live Demo: Messy to Clean Workflow
Input: Messy WhatsApp chat
Disorganized conversation with multiple topics
Prompt 1: "Summarize key updates and names"
Extract essential information
Prompt 2: "Create action items by person"
Organize responsibilities
4
Output: Clear report for boss/client
Professional, organized information
AI-Friendly Output Types
Bullet-point summaries
Checklists
FAQs
Email drafts
Decision support tables
Choose formats that organize information clearly and facilitate decision-making
Choose One Real Workflow
Select a document or tool flow you regularly use (e.g., Zoom → Summary Email)
1
1
Input Format
What format does your data start in?
Time Investment
How many hours per week does this take?
Pain Points
Where do you waste the most time?
4
4
Final Format
What's your desired end product?
Design a Prompt for Your Task
Role Definition
"You are a legal document specialist..."
Input Description
"Your input is a meeting transcript..."
Output Requirements
"Your output should be a bulleted summary..."
Style Guidelines
"Keep it concise and professional..."
Prompt Lab Activity
Create Variations
Design 2 different versions of your prompt
Test Both
Run them with identical input data
Compare Results
What changed? What improved? What surprised you?
Add Formatting to Prompts
Markdown
"Return this in markdown format"
Tables
"Give me a table with 2 columns: Action / Deadline"
Checkboxes
"Use checkboxes in the output"
Unstructured Data Prompt Examples
Chat Threads
"Summarize this WhatsApp thread by topic"
Freeform Notes
"Rewrite this freeform note as a formal brief"
Temporal Data
"Extract action verbs and dates from this text"
Peer Sharing: Before and After
Original Format
Tool → Text
Prompt Design
What instructions worked best?
3
3
New Output
The transformed result
Insights
Where did GPT struggle?
Mini-Sprint Briefing
Select
Choose 1 real workflow from your practice
Experiment
Write and test 3 different prompt versions
Evaluate
Assess output quality, time saved, and clarity
Before/After Template
Use this comprehensive template to document your AI workflow transformation process
Tool Used
What application contained your original data?
(e.g., Excel, Word, Email, Slack, Zoom)
Include version information and any relevant plugins or add-ons
Original Text
Brief description of input format and quality
Note formatting issues, length, and organizational challenges
Consider including a small sample excerpt for reference
Prompt Design
What instructions did you give the AI?
Include the full prompt text you used for best results
Note any iterations or refinements you made
GPT Output
What was produced and how useful was it?
Evaluate quality, accuracy, and time saved
Note any unexpected benefits or limitations
Efficiency Gains
Quantify time saved (minutes/hours per task)
Describe quality improvements compared to manual processing
Note any error reduction or consistency benefits
Next Steps
How could you further improve this workflow?
What other similar processes could benefit from this approach?
Consider automation opportunities to reduce manual steps
Complete this template for each workflow you transform.
Compare results with colleagues to identify patterns and best practices.
Key Takeaways & Reflection
Remember
  • AI needs clean, clear input
  • Text-exporting tools = easier wins
  • Structuring data is half the job
Reflect
  • What tool is most AI-ready?
  • Which messy task will you tackle?
WhatsApp: wa.me/6598419481