Week 4: AI Intermediate

Prompt libraries and chaining for complex tasks

The 80/20 rule of AI prompting

80% of effective AI output comes from context and setup.

Only 20% comes from the actual question you ask. This is the opposite of how most founders use AI.

The typical founder mistake

Jumping straight to questions without providing context.

“Write me an email to a prospect.” Then wondering why the output is generic and useless.

The solution

Build reusable context blocks you paste before any question.

Create a library of prepared context. Stop starting from zero every time.

Essential components

  • Company name and stage
  • What you do
  • Target customers
  • Value proposition
  • Competitors
  • Current metrics

Additional context

  • Key metrics you track
  • Voice and tone
  • Current focus

This document gets pasted at the start of every business-related prompt. Update it monthly.

Communication templates

Prospect research: Transform 30 minutes of manual research into 5 minutes. Find pain points, personal connections, decision makers.

Cold outreach, follow-ups, social content: All personalized at scale.

Business decision templates

Structured thinking for important choices:

  • Hiring decisions
  • Pricing changes
  • Partnership evaluations
  • Product pivots

Get comprehensive pros and cons, hidden risks, alternative approaches.

Operations templates

SOP creation: Document any repeatable process with step-by-step instructions, quality checks, troubleshooting guidance.

Describe what happens. Let AI structure it properly.

Research templates

  • Customer discovery
  • Competitor analysis
  • Pricing research

Each template includes your business context automatically.

Introduction to prompt chaining

Single prompts work for simple questions.

Complex tasks need chaining - a sequence of prompts where each builds on previous output.

How chaining works

  • Each step uses output from previous step
  • You review and guide between steps
  • Final output is deeper and more accurate
  • You maintain control throughout

Think of it as collaborative work, not magic automation.

Customer feedback analysis

Step 1: Identify patterns from customer calls

Step 2: Rank by importance and intensity

Step 3: Generate actionable insights

Step 4: Create follow-up questions

Market positioning

  • Analyze competitors
  • Find gaps
  • Define differentiation
  • Create ideal customer profile
  • Identify channels
  • Design messaging

Six connected steps producing complete positioning strategy.

Why chaining beats single prompts

  • Prevents incorrect assumptions
  • Gives you control
  • Produces better thinking
  • Matches how humans actually work

MCP: Connecting AI to your tools

Model Context Protocol lets AI connect to your apps without manual copy-paste.

This is the next evolution beyond manual prompting.

The manual way

  • Open Gmail
  • Copy relevant emails
  • Paste into ChatGPT
  • Ask for analysis
  • Copy response
  • Paste back into Gmail

Slow, repetitive, friction-filled.

With MCP

Ask: “Summarize unread emails from prospects and draft responses”

AI connects to Gmail automatically, reads, analyzes, drafts, presents summary for your review.

MCP use cases

Gmail: Summarize customer support emails, draft responses, find emails about specific topics

Google Drive: Summarize documents, find inconsistencies, identify gaps

Slack, Notion, Calendar, Database: Direct access and intelligent analysis

AI image generation

Beyond text, AI generates visual content for marketing and social.

The basic formula

Subject + Style + Setting + Lighting + Mood + Technical specs + Brand elements

Example: “Modern workspace with laptop showing analytics dashboard, minimalist aesthetic, natural window light, professional atmosphere, shallow depth of field”

Why this matters

  • Generate images in seconds versus waiting days
  • Unlimited iterations until it matches vision
  • Perfect brand consistency
  • Dramatically cheaper than custom photography

Organizing your prompt library

Personal prompts: Voice profile, background, communication templates, content creation templates

Business prompts: Company context, research templates, operations templates, strategy templates

Storage and access

Keep in Google Doc, Notion database, or text file in cloud storage.

Copy relevant context before every AI conversation.

The more you use these templates, the faster your AI work becomes.

The compounding value

Building this library takes focused time upfront.

But each template gets reused dozens of times.

When 80% of your prompt is ready to paste, every AI interaction becomes 10x more effective.

The core takeaway

Stop writing prompts from scratch every time.

Build a library of reusable context and proven templates. Use chaining for complex tasks.

The founders who build great prompt libraries compound their AI advantage over time.

Workshop

What you will create:

  • Complete business context document
  • Personal section with communication templates
  • Business section with research and decision templates
  • Minimum 10 tested templates
  • One complete prompt chain solving a real problem

Questions?

Next week: AI Advanced - Agents and autonomous operations