Assessment Blueprint — Building AI Agents: Use Cases

Section A

Summary

12

Skills assessed (of 28 in blueprint)

10

Skills at Apply / Analyze or above

Create

Bloom's ceiling — M8 two-agent portfolio

4

Assessment methods — all four in blueprint

6

Standards frameworks covered

4

Primary domain skills — AI Systems & Agents, all score 3–4

Section B

Bloom's Coverage Distribution

1%

8%

Apply 40%

Analyze 32%

Eval 15%

4%

Remember 1%

Understand 8%

Apply 40%

Analyze 32%

Evaluate 15%

Create 4%

Highest Evaluate/Create ceiling of the three v2 courses. Agents require judgment about autonomy, risk, and oversight — Evaluate and Create naturally surface in debate modules and the M8 capstone.

Section C

Skills Coverage by Domain

AI Systems & Agents Primary

Skill	Score	Bloom's Ceiling	Assessment Methods
Design AI Agents & Orchestration design-ai-agents	4 / 4	Evaluate	Quiz Module Test Lab Project
Implement AI Workflows & Automation implement-ai-workflows	4 / 4	Analyze	Module Test Lab Project
Manage Agent Memory & Context manage-agent-memory-context	3 / 4	Analyze	Module Test Lab
Integrate Multiple AI Models integrate-multiple-ai-models	3 / 4	Analyze	Module Test Lab

Prompt Engineering Partial — agent-focused

Skill	Score	Bloom's Ceiling	Assessment Methods
Leverage System Prompts & Instructions leverage-system-prompts	4 / 4	Evaluate	Quiz Module Test Lab Project

Generative AI Partial

Skill	Score	Bloom's Ceiling	Assessment Methods
Use Generative AI Tools & APIs use-generative-ai-tools	3 / 4	Apply	Module Test Lab Project

AI Ethics & Governance Partial

Skill	Score	Bloom's Ceiling	Assessment Methods
Ensure Responsible AI Deployment ensure-responsible-ai	3 / 4	Evaluate	Module Test Lab
Understand AI Ethics & Responsible AI understand-ai-ethics	2 / 4	Analyze	Module Test Lab

AI Applications Partial

Skill	Score	Bloom's Ceiling	Assessment Methods
Evaluate & Compare AI Solutions evaluate-ai-solutions	3 / 4	Evaluate	Module Test Lab Project

AI Fundamentals Partial

Skill	Score	Bloom's Ceiling	Assessment Methods
Grasp LLM Behavior & Capabilities grasp-llm-behavior	2 / 4	Apply	Module Test
Recognize AI Limitations & Failure Modes recognize-ai-limitations	3 / 4	Analyze	Module Test Lab

AI Security Partial

Skill	Score	Bloom's Ceiling	Assessment Methods
Secure AI Systems & Implementations secure-ai-systems	2 / 4	Apply	Module Test Lab

Section D

Assessment Method Breakdown

Method 1

Lesson Quizzes

Short multiple-choice and true/false checks on core agent concepts, tool definitions, and system prompt structure.

M1 agent architecture · M2 system prompts · M4 tool definitions

Method 2

Module Test (exam.html)

30 standards-mapped questions covering agent design, workflow, failure modes, oversight, and safe action boundaries.

All 8 modules · Bloom's: Understand → Analyze

Method 3

Conversational Labs

Hands-on AI-led labs in every module: SKILL labs (~25 min), DEBATE labs (~20 min), and BUILD labs (~30 min).

All 8 modules · Bloom's: Apply → Evaluate

Method 4

Integration Project

M8 "Your Agent Portfolio" capstone: design, build, document, and evaluate two production-ready agents.

M8 capstone · Bloom's: Evaluate → Create

Section E

Standards Alignment

Standard / Framework	Specific Articles or References	Modules
AI4K12	Big Idea 5 (Society & Ethics), Big Idea 2 (Representation & Reasoning)	M1, M3, M7
EU AI Act	Article 14 (human oversight), Article 22 (automated decisions), Article 9 (risk management)	M3, M7, M8
NIST AI RMF	Govern, Map, Manage, Measure — all four RMF functions	M3, M7, M8
O*NET	Technology Skills, Systems Analysis, Programming	M1, M2, M4, M5, M6, M8
ISTE	Computational Thinker 5c, Innovative Designer 4c	M1, M4, M5, M6, M8
WEF Future of Jobs	AI and big data, Analytical thinking	M1, M3, M7

Section F

Gaps & Recommendations

Intentional Scope Boundaries

This course does not cover RAG / retrieval-augmented generation, vector databases, or knowledge graph design — those belong to a dedicated knowledge-systems course. Fine-tuning and model training are out of scope. Data analytics and predictive modeling are not addressed.

The ethics and governance treatment (M3, M7) is deep enough to inform responsible agent deployment, but it is not a primary objective. Learners who need comprehensive governance coverage should pair this course with AI Ethics & Decision Making.

For learners new to prompting, completing Building with AI first is recommended — that course builds the prompt-engineering foundation (write-clear-prompts, design-multi-turn-conversations) that M2 of this course assumes.

Section G

Portfolio Artifact

Capstone Deliverable · M8

Your Agent Portfolio

Two production-ready AI agents with documented architecture, tool definitions, safety constraints, and evaluation criteria. Each agent submission includes a system prompt, at minimum two integrated tools, a stated autonomy boundary, and a written evaluation against defined success criteria.

Standards mapped: EU AI Act Articles 14 and 22 · NIST AI RMF Govern and Manage functions · O*NET Systems Analysis · ISTE Innovative Designer 4c