This section teaches the decisions that come before implementation. The sequence begins with the simplest automation options, moves through system and data readiness, and ends with evaluation, architecture choice, and economics.
What You Will Be Able to Do
- Choose the simplest adequate approach instead of defaulting to an LLM or agent.
- Separate prompting, retrieval, fine-tuning, workflows, and agent-like behavior.
- Evaluate data readiness, failure modes, business value, cost, latency, and operating fit.
- Build a fixed evaluation set before comparing systems.
Guided Sequence
Decision foundations
- Rules, RPA, ML, LLMs, and Agents: The Decision Ladder — A technology-choice record selecting the simplest adequate automation approach.
- The AI Stack in Plain English — A system map identifying the model, context, tools, controls, and operating owner.
- What AI Gets Wrong — A failure-mode register mapping likely errors to detection and containment controls.
- How to Evaluate an AI Use Case — A scored AI use-case assessment with an owner, success measure, and stop/go decision.
- Data Readiness for AI Projects — A data-readiness inventory with owners, access conditions, quality risks, representativeness gaps, and remediation actions.
Method selection and evaluation
- LLMs vs Traditional Machine Learning — A justified model-family decision for one business problem.
- Prompting, RAG, Fine-Tuning, Workflows, or Agents? — A method-choice record explaining the selected pattern, rejected alternatives, evidence needed, and review boundary.
- Prompting 101 for Business — A reusable prompt specification with inputs, constraints, output schema, and review criteria.
- How to Build an AI Evaluation Set — A versioned evaluation pack with representative cases, expected outputs, scoring rules, and release thresholds.
- RAG Explained for Business — A retrieval-system decision specifying sources, citations, permissions, and abstention.
- AI Agents vs Workflows — An autonomy decision that states whether the use case needs a fixed workflow or agent-like behavior.
Economics and background
- Cost, Latency, and ROI of AI Systems — A unit-economics model covering inference, latency, review effort, and business value.
- AI Timeline of Modern AI — An evidence-based timeline connecting major AI shifts to current business capabilities.
- Where to Go Deeper Beyond This Academy — A goal-specific advanced study plan using primary papers and technical references.
Lesson Library
| Lesson | Level | Time | Learner output |
|---|---|---|---|
| Rules, RPA, ML, LLMs, and Agents: The Decision Ladder | Intermediate | 20 min | A technology-choice record selecting the simplest adequate automation approach. |
| The AI Stack in Plain English | Beginner | 15 min | A system map identifying the model, context, tools, controls, and operating owner. |
| What AI Gets Wrong | Beginner | 20 min | A failure-mode register mapping likely errors to detection and containment controls. |
| How to Evaluate an AI Use Case | Beginner | 20 min | A scored AI use-case assessment with an owner, success measure, and stop/go decision. |
| Data Readiness for AI Projects | Beginner | 20 min | A data-readiness inventory with owners, access conditions, quality risks, representativeness gaps, and remediation actions. |
| LLMs vs Traditional Machine Learning | Beginner | 20 min | A justified model-family decision for one business problem. |
| Prompting, RAG, Fine-Tuning, Workflows, or Agents? | Intermediate | 20 min | A method-choice record explaining the selected pattern, rejected alternatives, evidence needed, and review boundary. |
| Prompting 101 for Business | Beginner | 20 min | A reusable prompt specification with inputs, constraints, output schema, and review criteria. |
| How to Build an AI Evaluation Set | Intermediate | 25 min | A versioned evaluation pack with representative cases, expected outputs, scoring rules, and release thresholds. |
| RAG Explained for Business | Intermediate | 20 min | A retrieval-system decision specifying sources, citations, permissions, and abstention. |
| AI Agents vs Workflows | Beginner | 20 min | An autonomy decision that states whether the use case needs a fixed workflow or agent-like behavior. |
| Cost, Latency, and ROI of AI Systems | Intermediate | 15 min | A unit-economics model covering inference, latency, review effort, and business value. |
| AI Timeline of Modern AI | Beginner | 20 min | An evidence-based timeline connecting major AI shifts to current business capabilities. |
| Where to Go Deeper Beyond This Academy | Advanced | 25 min | A goal-specific advanced study plan using primary papers and technical references. |
Completion Standard
A lesson is complete when the stated learner output has been produced, reviewed against representative evidence, and assigned an owner or next decision. Reading the page without producing the artifact is orientation, not completion.
Practice Case
Use the Harborline Services running case when you do not have safely redacted examples from your own organization.
Where to Go Next
- Return to the Academy home
- Browse the Academy Practice Cases