I Built a $20,000 AI Consultant You Can Have For Free

March 12, 2026

21 min episode · 2 min read

Episode

21 min

Read time

2 min

Topics

Artificial Intelligence

AI-Generated Summary

Published Mar 15, 2026

Key Takeaways

✓AI Skill Files as Workflow Analyzers: Upload a screen recording transcript alongside a skill file (markdown format) to Claude, ChatGPT, or Perplexity to generate a task transformation report. The skill auto-scopes output length by work type: 2–4 pages for single tasks, up to 40 pages for department-level transformations, keeping recommendations proportionate to complexity.
✓Time Savings Estimation Method: The AI transformation skill calculates current task duration from the recording itself — a 2-minute narrated walkthrough yields estimated times, while a 3-hour working session produces exact figures. For a basic slide review task, the skill estimated recovery of 1–2.5 hours per week for a CMO-level role.
✓Meta Skill for Building Skills: A second "skill architect" file, built by crawling GitHub and other skill directories for best practices, generates new skill files following proven architecture patterns. It handles single-file versus multi-file decisions, skill families, and cross-platform compatibility, all within the 500-line maximum constraint that governs AI skill files.
✓AI Adoption Accelerates Where New Capabilities Unlock: Behavioral change toward AI tools happens faster when people use AI to do things they previously could not do at all — coding, data analysis, complex research — rather than replacing existing habits. Starting with small, novel use cases builds momentum before tackling workflow replacement.
✓High Agency plus Low Tolerance as the Adoption Framework: The people making the fastest AI transitions share two traits: high agency (proactively building and experimenting) and low tolerance for inefficiency (unwilling to accept slow processes when AI can fix them in minutes). Identifying these people on a team predicts who will lead successful AI transformation.

What It Covers

Kipp Bodnar demonstrates a free AI transformation skill file built in Claude that analyzes screen recordings of any work task and generates a structured report showing current versus AI-assisted workflows, estimated time savings, tool recommendations, and step-by-step implementation plans for individuals and teams.

Key Questions Answered

•AI Skill Files as Workflow Analyzers: Upload a screen recording transcript alongside a skill file (markdown format) to Claude, ChatGPT, or Perplexity to generate a task transformation report. The skill auto-scopes output length by work type: 2–4 pages for single tasks, up to 40 pages for department-level transformations, keeping recommendations proportionate to complexity.
•Time Savings Estimation Method: The AI transformation skill calculates current task duration from the recording itself — a 2-minute narrated walkthrough yields estimated times, while a 3-hour working session produces exact figures. For a basic slide review task, the skill estimated recovery of 1–2.5 hours per week for a CMO-level role.
•Meta Skill for Building Skills: A second "skill architect" file, built by crawling GitHub and other skill directories for best practices, generates new skill files following proven architecture patterns. It handles single-file versus multi-file decisions, skill families, and cross-platform compatibility, all within the 500-line maximum constraint that governs AI skill files.
•AI Adoption Accelerates Where New Capabilities Unlock: Behavioral change toward AI tools happens faster when people use AI to do things they previously could not do at all — coding, data analysis, complex research — rather than replacing existing habits. Starting with small, novel use cases builds momentum before tackling workflow replacement.
•High Agency plus Low Tolerance as the Adoption Framework: The people making the fastest AI transitions share two traits: high agency (proactively building and experimenting) and low tolerance for inefficiency (unwilling to accept slow processes when AI can fix them in minutes). Identifying these people on a team predicts who will lead successful AI transformation.

Notable Moment

Kieran pushes back on the detailed report format, arguing a 4–6 page output for a single workflow feels overwhelming. Kipp reframes it: the report is not meant to be read linearly but fed directly back into an AI tool to auto-generate the specific prompts and agent builds needed.

Know someone who'd find this useful?