Skip to content
Nerdy beta

Effective AI

Flow AI

Your team already uses AI: chatting, generating drafts, asking questions. What you don’t have is a compounding return on it. This workshop takes your team past casual usage into deliberate workflow engineering — finding the repetitive, toilsome tasks that eat the week, then engineering AI to handle them so your people spend their attention on the judgment, creativity, and relationships no model can replace.

The workshop is hands-on and grounded in your team’s actual work. We begin with fundamentals: navigating the Claude ecosystem, writing effective prompts, and supplying the right context. From there, we move into compounding engineering: examining the workflows your team runs every week, breaking them into discrete tasks, and building reusable AI setups that get better over time. Every exercise produces something immediately useful. Teams leave with working use cases and an action plan for finding more value without us in the room.

Who It’s For

Teams of casual AI users ready to become intentional engineers of their own workflows. Whether your people have dabbled with AI tools or use them daily without a system, the workshop starts where they are. Cohorts accommodate up to 20 people.

Format

  • Remote: 4 sessions delivered over 2-3 weeks
  • In-person: 1 day

What You’ll Walk Away With

  • Practiced fluency navigating the Claude ecosystem (Desktop, browser, Co-work, Code) and selecting the right modality for different tasks
  • Effective prompting skills using prompt engineering, reverse prompt engineering, meta-prompt engineering, and context engineering techniques
  • A “use case map” for a recurring team workflow, decomposed into AI-augmentable tasks using metacognitive workflow analysis
  • At least one working Claude Project built around a real, recurring task
  • A set of guiding principles for effective AI usage and a recommended action plan for continued learning
  • A digital whiteboard with all workshop curriculum, examples, exercises, and templates

Example Sessions

Session 1: Getting in the Water — The Claude Ecosystem & Finding Use Cases

We start with the case for AI fluency and the opportunity ahead. Teams learn to navigate the Claude ecosystem: Desktop, browser, Co-work, and Code, understanding what each is for and when to reach for it. We cover Claude basics like Plan Mode, environment setup, and care and feeding. The session introduces how to curate use cases that deliver real value, not just more docs and markdown, and challenges people to separate their job’s purpose from its tasks in order to let go of entrenched habits. The session closes with an exercise: take a real task, try it with Claude, and share what happened.

Session 2: Thinking Like an Engineer — Prompt Engineering & Context Engineering

This session builds the core technical skills. We cover prompt engineering fundamentals (structure, specificity, iteration), introduce named prompt frameworks like CO-STAR and STOKE as scaffolding for completeness, then move into reverse prompt engineering and meta-prompt engineering: using AI to sharpen the prompts themselves. Context engineering teaches how to supply the right context the right way. We walk through the anatomy of a great prompt and explore when less is more versus when detail pays off. Teams craft effective prompts for real tasks, then swap and peer-review each other’s work.

Session 3: Compounding Engineering — Intentional Workflow Design

The pivot from prompting to systems thinking. Teams learn metacognitive workflow analysis: thinking deliberately about how the work actually gets done. We use lightweight value stream mapping to identify steps, handoffs, and bottlenecks in recurring workflows, then decompose those workflows into AI-augmentable tasks. The session covers the shift from tacit to explicit knowledge and teaches compounding engineering: building Claude Projects, custom instructions, and example-driven setups that improve with use. Teams map a recurring workflow and identify the highest-value AI insertion points.

Session 4: Path Forward — Putting It All Together

We close by addressing where humans stay in the loop: quality, judgment, and preventing workslop. Teams apply everything they’ve learned to a real, meaty task end-to-end. The session finishes with each team developing an action plan: 3-5 principles for effective AI usage and the next use cases to pursue. We discuss what sustains momentum and where to find support and resources for continued growth.

  • Value Stream Mapping — used in Session 3 to map recurring workflows
  • Prompt Frameworks — six named prompt structures (CO-STAR, RISEN, STOKE, and more) with a decision matrix, introduced in Session 2
  • Compounding Engineering — the core practice of building AI setups that improve over time
  • Metacognition — the skill of examining work patterns deliberately
  • mcp-server — connect your AI tools directly to this knowledge collection