Introduction

Struggling to make AI actually useful in your daily work? This guide shows how to turn an LLM into your smartest teammate.

What’s the problem it solves?

Many people try AI tools but quickly drop them because the results feel vague or unhelpful. The problem isn’t the AI-it’s the lack of context. This guide solves that by teaching you how to give AI the same background you'd give a new coworker, making it way more helpful.

Quick Summary

Tal Raviv shares how he went from ignoring AI to using it daily as a trusted partner. He explains that most people give up on AI because they don’t provide enough context or direction. But when AI has the same info you’d give a real teammate-your goals, company background, team dynamics-it becomes a powerful sidekick that helps you think better, not just faster.

He breaks the process into four steps: “hire” your copilot (set its values and behaviors), “onboard” it (upload strategy docs, past feedback, etc.), start a project thread (share everything you know about your current task), and then ask it to help you decide what to do next. The more you work with it, the more valuable it becomes-remembering your lessons, nudging you with ideas, and helping you move faster with less stress.

Beyond the basics, Tal gives advanced use cases like building prototypes, running AI automations, and even just venting when things go sideways. He ends by imagining a future where copilots are team-aware, push us to act, and update themselves automatically-all so we can work smarter, not harder.

Key Takeaways

• Your AI is only as smart as the context you give it
• Don’t expect magic from a cold prompt-treat your AI like a teammate
• Use projects or threads to separate your initiatives and keep context clean
• AI copilots can help with deep thinking, not just shallow tasks
• “Gossiping” to your AI helps it stay updated like a real coworker
• Building habits around your AI copilot is more important than using fancy tools
• Future copilots may become shared team knowledge hubs and proactive assistants

What to do

• Choose the LLM with the best reasoning ability available to you
• Write a custom “instruction prompt” that defines your copilot’s role and tone
• Upload docs about your company, product, and personal goals to give it context
• Start a new thread for each work initiative and “dump your brain” to get started
• Regularly update the AI with key changes, lessons, or even frustrations
• Use your copilot for strategy, automation ideas, prototyping, and personal development
• Build the habit: keep the AI open like you would Slack or Notion
• When threads get too long, ask it to summarize and restart a fresh thread with retained memory
• Invest in growing its knowledge base over time-small inputs now pay off later

‍