
| Jobs created by AI by 2030 | 170 million (author's claim) |
| Work done by agents in Martell's company | 92% (author's claim) |
| Performance drop when airline agent lost its rulebook | 33% to 11% (shown on screen) |
| Cost of refactoring codebase with Opus | $150 (shown on screen) |
| Cost of refactoring codebase with Haiku | $1.50 (shown on screen) |
Who will this analysis save an hour?We’ll send the link with the “actually” takeaway — they won’t have to watch the whole video.
AI agents are autonomous systems designed to handle full workflows rather than just answering questions. They can diagnose a situation, build a plan, take action in external systems, and assess their own work to improve over time.
Start by defining a clear outcome and a 'Definition of Done'. Then, create three identity files: a Soul file for personality/voice, an Identity file for the job description, and a User file for context about who the agent is serving.
The Rule of R helps you decide what to automate. A task is a good candidate if it is Repetitive (done frequently), Rule-based (consistent input/output), and provides a high Return on the time invested to build it.
Trust should be built in stages. Set guardrails in the identity files, have the agent 'approve everything' through you first, then slowly loosen the leash as it proves accuracy, and finally set it on a recurring schedule (heartbeat).
★ 9/10
★ 9
★ 9