The Next Input — Issue #106

The AI "Wrapped" You Actually Need

Author

Aaron Bost
December 23, 2025

In partnership with

Mariah Carey Christmas GIF by Justin Bieber

⚡ The Briefing — 60 sec

🛠️ The Playbook — The Personal AI Retrospective Engine

Mission Turn your AI usage into a clear, honest annual review—what you used, what actually worked, and what turned into leverage.
Difficulty Medium
Build time 2 hours
ROI Turns vague “AI helped me” vibes into concrete insight you can double down on next year.

0) Why This Matters

Wrapped-style summaries are fun—but shallow.
If AI actually moved the needle for you this year, you should know how, where, and why.
This engine creates a real retrospective: usage patterns, outcomes, and the moments where AI crossed from novelty into utility.

1) Architecture

Component

Tool

Purpose

Data Intake

Chat logs / exports

Raw interaction history

Categoriser

Claude 4.5 Haiku

Group usage by intent

Impact Analyzer

GPT-5-mini

Identify outcomes + leverage

Memory Store

Notion / Airtable

Structured yearly record

Report Layer

Markdown / PDF

Shareable year-in-review

2) Workflow

  1. Export or collect AI usage data (chat logs, prompts, tasks).

  2. Claude 4.5 Haiku categorises usage:

    • learning

    • execution

    • automation

    • ideation

    • revenue-related work

  3. GPT-5-mini analyses patterns:

    • time saved

    • complexity handled

    • moments of compounding value

  4. System flags “breakthrough moments” where AI changed trajectory.

  5. Generate a clean Year-in-Review report.

  6. Use insights to decide what to double down on next year.

3) Example Prompts

Usage Categorisation (Claude 4.5 Haiku)

Categorise these AI interactions by intent:
- learning
- execution
- automation
- ideation
- revenue or business creation
Return totals and notable spikes.

Impact Analysis (GPT-5-mini)

Analyse these categories to identify:
- where AI saved the most time
- where it created new capability
- any compounding effects
Summarise in clear business language.

4) Guardrails

  • Don’t inflate impact—stick to evidence.

  • Separate experimentation from real outcomes.

  • Avoid vanity metrics (message count ≠ value).

  • Keep the review private unless you choose to share.

5) Pilot Rollout — 2 hours

  1. Collect three months of AI usage as a test.

  2. Run categorisation and impact analysis.

  3. Manually validate one week for accuracy.

  4. Generate a mini-report.

  5. Expand to full-year data.

  6. Lock in as an annual ritual.

6) Metrics

  • Hours saved across the year

  • Tasks automated vs manual

  • Revenue or output influenced by AI

  • Skill areas accelerated

  • Compounding workflows identified

Pro Tip: The real value isn’t the summary—it’s spotting the one workflow that quietly changed everything.

🎯 The Arsenal — Tools & Platforms

Copy-paste prompt block:

Analyse my AI usage this year.
Identify:
- what actually mattered
- what didn’t
- where leverage appeared
Be honest and practical.

💡 Free Office Hours

Want help implementing anything? Book a free 15-minute Office Hours slot—no sales pitch, just workflows solved.

Get the 90-day roadmap to a $10k/month newsletter

Creators and founders like you are being told to “build a personal brand” to generate revenue but…

1/ You can be shadowbanned overnight
2/ Only 10% of your followers see your posts

Meanwhile, you can write 1 email that books dozens of sales calls and sells high-ticket ($1,000+ digital products).

After working with 50+ entrepreneurs doing $1M/yr+ with newsletters, we made a 5-day email course on building a profitable newsletter that sells ads, products, and services.

Normally $97, it’s 100% free for 24H.

🕹️ Game Over

Reflection turns tools into leverage.

— Aaron Automating the boring. Amplifying the brilliant.