Six weeks. Twelve to fifteen people. One outcome: a team that can direct any AI system, on any platform, against your real work.
What it is
A six-week, in-person and live-virtual program that takes a team from "we use ChatGPT sometimes" to "we use AI systematically against our actual workflows, and we know why we're getting the results we're getting."
Built around computational thinking — the cognitive scaffolding that makes AI fluency durable. Each week introduces a new principle (decomposition, prompt design, pattern recognition, multimodal integration, data analysis, strategic implementation) and applies it to participants' real work. No hypothetical exercises. No generic case studies.
How a week runs
Every session is the same 90-minute shape:
| Block | Time | What happens |
|---|---|---|
| Pre-read + homework review | (async + 10 min) | Pre-reading lands 3–5 days before; session opens with homework debrief |
| Core concepts & capabilities | 25 min | The technical work — prompting, evaluation, workflow patterns — for the week |
| Foundations & mental models | 25 min | The cognitive principle that underlies the technical work — analogies, worked examples, the "why this works" layer |
| Hands-on application | 25 min | Live work on your actual workflows. Pair work in groups of 3–5 |
| Debrief + next assignment | 5 min | Pattern naming, retrieval prep, preview |
Between sessions: instructor reviews and gives feedback on homework. Slack/Teams channel stays open. Optional office hours.
The six weeks
| Wk | Core Concepts & Capabilities | Foundations & Mental Models | Hands-On Component |
|---|---|---|---|
| 1 | How Large Language Models work through pattern recognition, tokenization, and embeddings | Grounding with introduction to computational thinking as foundation for AI understanding | Overview of existing AI tools and models with focus on selecting the right tool for specific business tasks |
| 2 | Prompt engineering as algorithm design with precise instructions and structure | Decomposition principles for creating effective step-by-step AI instructions | Building algorithmic prompts and developing debugging approaches |
| 3 | Advanced prompting techniques as systematic patterns | Abstraction, iteration, and pattern recognition for refining AI interactions | Creating evaluation frameworks and testing to optimize prompts |
| 4 | Applying text-based patterns to image generation and multi-step workflows | Orchestration across different domains and connected workflow systems | Building cross-modal systems integrating multiple AI capabilities |
| 5 | Structured analysis processes transforming data into insights | Computational approaches to data analysis through abstraction and modeling | Decomposing complex data problems and building practical applications |
| 6 | Workflow development and coding assistance | Systems thinking and integration of principles across organizational processes | Creating implementation roadmaps and systematic evaluation approaches |
What participants walk out with
- Skills that transfer. Computational thinking applied to AI — survives every model upgrade, every new tool, every platform change. We are not training on a specific tool.
- Top-tier prompting capability. Demonstrated against the team's real work, not synthetic exercises.
- Agent design instincts. They can structure problems clearly enough to delegate to AI and know when to verify versus trust.
- A reusable prompt library built during the program, owned by your organization.
- A named pilot working team. The five evidence-supported computational-thinking pillars surface high-performers naturally. By week 6 we know who should lead the next wave of internal AI work.
What the organization walks out with
- A documented baseline (pre-program) and outcome (post-program) assessment.
- A capability assessment that names champions and identifies gaps.
- Reusable assets: prompt libraries, exercise templates, evaluation rubrics, reference materials.
- An implementation roadmap built collaboratively in week 6.
- A specific, written plan for what next 6 months should look like.
Why it works
Three reasons, each anchored in research, all visible in client outcomes:
1. Computational thinking is the transfer mechanism. Decades of cognitive-science research (Salomon & Perkins 1989, Schwartz & Bransford 1998, the K-12 CT meta-analyses) converge on one finding: what transfers is abstract schema, not procedural keystrokes. A program that teaches the tool will produce tool-dependent users. A program that teaches the underlying principles produces fluent users. Our pedagogy is built explicitly around the latter.
2. Customization is structural, not cosmetic. Pre-program we review your existing workflows and tailor every exercise to them. Client data flows into the hands-on segments when available. Department-specific examples replace generic ones. The methodology guide names "customization to organizational context" as one of five non-negotiable pillars.
3. Measurement is honest. We commit, in writing, to specific success criteria before the program runs. Pre/post assessment scores, homework rubric quality, use-case identification, NPS, and post-program organizational impact indicators (processes implemented, time/cost saved, cross-functional collaboration instances). If those numbers don't move, we say so and recommend a redesign.
Results from the most recent published cohort (Une Femme Wines, 2025)
- 100% consistent daily AI usage in 6 weeks (from 41%)
- +165% in participants reporting transformational impact (17% → 45%)
- +350% code-writing capability, +156% summarization, +148% research, +121% data analysis
- ~4.4 hours/week saved per person
- 8.5/10 NPS, 7.5/10 confidence using LLMs daily, 8.1/10 interest in continued learning
"Happy Robots taught us how to think about generative AI before pushing tools on us. That foundation made all the difference." — Jen Pelka, co-founder
"We went from people experimenting with ChatGPT to having an AI-forward organization in six weeks. That's not about the hours saved — it's about what we can build from here." — Thomas Hartman, VP Operations
What we ask of you
- 12–15 participants split into working groups of 3–5
- 2 hours/week per participant (90-min session + ~30 min homework)
- Access to representative real workflows — we coach on your work, not ours
- Leadership sponsorship through week 6, including a closing roadmap session