My AI Productivity Stack: The Complete Workflow

People ask me all the time: "What AI tools do you actually use?"

Fair question. There's a lot of hype and not enough practical guidance. Here's my actual setup—the tools I pay for, how I use them throughout the day, and the workflows that tie everything together.

This isn't aspirational. This is what I actually do.

My Core Stack

Daily Drivers

Total: ~$74/month

Is that a lot? Maybe. But I estimate these tools save me 10-15 hours per week. Do the math on your hourly rate.

Additional Tools (As Needed)

• Gamma (free tier): Quick presentations
• Otter.ai (free tier): Meeting transcription
• Canva (free tier): Quick design work
• Leonardo AI (free tier): Image generation for content

My Morning Routine (7-9 AM)

1. Inbox Triage with AI

I start by scanning overnight emails. For anything requiring a response:

Quick replies: I draft in my email client

Complex replies: I open Claude and prompt:

"Here's an email I received: [paste]. My situation is [context]. Draft a response that [goal]."

I rarely use the draft word-for-word, but it gets me 80% there.

2. Daily Planning

In Todoist, I review my tasks. But I use Claude to help prioritize:

"Here are my tasks for today: [list]. I have about 5 focused hours. Given that [context/priorities], what order should I tackle these?"

This takes 2 minutes and often catches blind spots in my own planning.

3. Research Queue

If I have articles to write or decisions to make, I queue up research in Perplexity:

• Open multiple tabs with questions
• Let them load while I do other things
• Come back to synthesized research

Deep Work Block (9 AM - 12 PM)

Writing Projects

My writing workflow is:

Phase 1: Research (Perplexity)

"[Topic] - what are the key debates? What are people missing?"

I read through the sources, not just the summary.

Phase 2: Outline (Claude)

"I'm writing about [topic] for [audience]. Here are my scattered thoughts: [notes]. Help me structure these into a logical outline."

Phase 3: Draft (Mostly me, some Claude)

I write the first draft myself. The voice comes from me, not AI. But for sections where I'm stuck:

"I want to make the point that [X]. Give me three ways to phrase this opening."

Phase 4: Edit (Claude)

"Review this draft. Identify: weak sections, unclear sentences, and any clichés. Don't rewrite—just point out problems."

I fix the problems myself. This maintains my voice while benefiting from AI-powered editing.

Coding Projects

I keep Claude open in a side pane while coding:

• Stuck on approach: "I want to [goal]. What are my options?"
• Debugging: "Getting this error: [error]. Here's the relevant code: [code]"
• Code review: "Review this function for efficiency and readability"

For longer coding sessions, Claude Code (terminal-based) is even better—it can see my entire codebase.

Meetings (Throughout Day)

Before Meetings: Prep

"I'm meeting with [who] to discuss [topic]. What questions should I ask? What should I prepare?"

"Summarize what I know about [company/person] for a meeting in 10 minutes."

During Meetings: Otter.ai

I run Otter for automatic transcription. This lets me focus on the conversation instead of taking notes.

After Meetings: Processing

I paste Otter transcripts into Claude:

"Here's a meeting transcript. Extract: 1) Key decisions, 2) Action items with owners, 3) Open questions we didn't resolve."

This goes into Notion as the meeting record.

Afternoon Block (1-5 PM)

Admin Tasks

AI handles the tedious stuff:

Email drafts: Complex responses, follow-ups, introductions

Documentation: I describe a process verbally, Claude turns it into proper documentation

Spreadsheet work: "Write a formula that [does X]" or "Analyze this data and tell me [Y]"

Creative Work

When I need ideas:

"I'm working on [project]. Give me 10 unconventional approaches I might not have considered."

I never use AI ideas directly, but they spark my own thinking.

Research Deep Dives

For complex research, I use Perplexity's Pro Search:

"I need to understand [complex topic]. Give me a comprehensive overview with sources. Then I'll ask follow-up questions."

The conversation continues as I go deeper. All sources are cited, so I can verify anything important.

End of Day (5-6 PM)

Daily Review

This is where Notion + AI shines:

"Here's what I accomplished today: [list]. Here's what's still open: [list]. What should I prioritize tomorrow?"

Inbox Zero Push

Any remaining emails get the AI treatment:

• Quick responses drafted
• Complex items flagged for tomorrow
• FYIs archived

Weekly Workflows

Monday: Planning

I paste my weekly goals into Claude:

"Here are my goals for this week. Break each into specific tasks I can accomplish each day."

Friday: Review

"Here's what I accomplished this week vs. what I planned. What patterns do you notice? What should I do differently?"

Ongoing: Learning

When I'm learning something new:

1. Perplexity for initial research
2. NotebookLM with my notes for deeper understanding
3. Claude for practice questions and explanations

The Workflows That Changed Everything

1. The Research-to-Draft Pipeline

1. Research in Perplexity → export as markdown
2. Import into Claude with my angle
3. Claude helps structure
4. I write the draft
5. Claude reviews for issues
6. Final edit is me

Time from blank page to finished draft: cut by ~40%

2. The Meeting Efficiency Loop

1. Prep questions with AI before
2. Otter transcribes during
3. Claude summarizes after
4. Actions go straight to Todoist

I never scramble after meetings anymore.

3. The Decision Documentation System

For any significant decision:

1. Capture options in Claude
2. Get pros/cons analysis
3. Document rationale in Notion
4. AI helps predict potential issues

Makes defending (and reviewing) decisions much easier.

What Doesn't Work

To be honest:

AI for creative conception: The big ideas still need to come from me

Full delegation: Everything needs review. AI makes mistakes.

Replacing thinking: Using AI for every small task actually slows you down

Non-stop AI: Sometimes just writing/thinking without AI is faster

My Advice for Building Your Stack

1. Start with one tool (I'd pick Claude Pro)
2. Use it for one workflow (I'd start with writing)
3. Get good at prompting before expanding
4. Add tools when you hit limits, not because they exist
5. Measure actual time saved, not theoretical

The goal isn't to use more AI. It's to get more done.

Monthly Cost vs. Value

My $74/month in AI tools replaces:

• Research assistant: ~$15/hour
• Editor: ~$50/hour
• Admin support: ~$25/hour

At 10 hours saved per week, the ROI is absurd. Even 2-3 hours would pay for it.

The question isn't "can I afford AI tools?" It's "can I afford not to use them?"

What does your AI stack look like? I'm always looking to optimize.