ARCH2360 — Assignment 2 Briefing 

AI Workflow Exercise

+ Process Visualization & System Thinking

Opening Briefing

“Alright everyone…”

“Today we are moving into Assignment 2.”

“And this assignment is very important…”

because now…
we move beyond:

‘using AI tools’

toward:

‘designing AI-assisted workflows.’

“Last assignment…”

you explored AI individually.

You tested:

• prompts,

• outputs,

• reflections,

• different AI personalities.

That was your exploration phase.

Now…

you must start thinking like architects.

And architects…

do not work with one tool only.

Architects orchestrate systems.

🎼

What Is Assignment 2 About?

“This assignment is called:

AI Workflow Exercise

• Process Visualization & System Thinking.”

Weightage:
20%.

Can be:

• individual,
  or

• pair work,
  depending on lecturer instruction.

The Big Idea

“In real architectural practice…”

you never use:

• one software,

• one platform,

• one AI only.

Instead…

your workflow may involve:

• BIM,

• rendering,

• research,

• climate analysis,

• presentation tools,

• AI assistants,

• simulations,

• diagrams,

• reporting systems.

Everything connected.

So this assignment trains you to think:

systematically.

Not randomly.

Core Question

The core question is:

“How can AI support architectural thinking and workflow?”

Not:

“How many AI tools can I use?”

That is different.

Very different.

Your Task

“You must choose ONE architectural scenario.”

You don’t need to do everything.

Choose ONE.

Examples already given:

OPTION A — Concept Design Workflow

Example flow:

Prompt
→ AI concept
→ image generation
→ diagrams
→ presentation

Example Exploration

You can explore:

• tropical housing

• futuristic library

• adaptive mosque

• student housing

• floating architecture

• AI-assisted façade concepts

Possible Workflow

ChatGPT
→ concept narrative

Midjourney
→ visual exploration

Canva AI / Gamma
→ presentation layout

Google Slides
→ final presentation

OPTION B — Site Analysis Workflow

Example:

Research AI
→ mapping
→ climate analysis
→ summary
→ design response

Example Topics

• flood-prone site

• urban heat island

• Klang Valley traffic

• coastal resilience

• walkability studies

Possible AI Tools

Perplexity
→ research

Gemini
→ summarize climate data

Google Maps / GIS
→ mapping

ChatGPT
→ interpretation

Then:
YOU decide the architectural response.

OPTION C — Building Performance Workflow

Example:

Model
→ simulation
→ AI analysis
→ interpretation
→ design refinement

Example Topics

• daylight analysis

• passive cooling

• ventilation

• energy optimization

• façade performance

Possible Tools

Revit
→ base model

Ladybug / Honeybee
→ simulation

ChatGPT
→ interpretation

Canva AI
→ workflow presentation

OPTION D — Construction / BIM Workflow

Example:

BIM
→ clash detection
→ AI monitoring
→ reporting
→ coordination

Example Topics

• site monitoring

• construction delays

• safety systems

• BIM coordination

• smart construction

Example Workflow

Revit
→ BIM model

Navisworks
→ clash detection

AI dashboard
→ issue analysis

Gemini / ChatGPT
→ reporting summary

OPTION E — Urban / Smart City Workflow

Example:

Urban data
→ AI analysis
→ visualization
→ planning proposal

Example Topics

• smart kampung

• AI traffic systems

• flood intelligence

• urban mobility

• sustainable districts

• smart public spaces

Example Workflow

Perplexity
→ urban research

Gemini
→ summarize trends

Midjourney
→ visual concepts

ChatGPT
→ urban strategy narrative

What Must You Show?

This is VERY important.

I do not only want final images.

I want PROCESS.

Architects must learn:

• process thinking,

• system logic,

• refinement,

• decision-making.

You Must Show

1. Workflow Sequence

Show:

• what happened first,

• second,

• third,

• fourth.

Like a system map.

2. AI Tools Used

Explain:

• why this tool?

• what role did it play?

Not random usage.

Purposeful usage.

3. Prompt Strategy

Show:

• original prompts,

• refinement prompts,

• contextual prompts.

And explain:

Why did your prompts evolve?

Because prompting itself is design thinking.

Example

Weak prompt:

“Design modern building.”

Very general.

AI confused.

Better prompt:

“Generate tropical passive cooling student housing for Shah Alam with shaded courtyard and natural ventilation inspired by Malay architecture.”

Now AI has:

• context,

• climate,

• identity,

• intention.

That changes the output dramatically.

Human Judgment

This is the most important part.

You must explain:

• What worked?

• What failed?

• Which AI was useful?

• Which AI was weak?

• Where did human judgment become necessary?

Because:

AI assists workflow.
Architects remain responsible for decisions.

Reflection Layer

This assignment is not only technical.

It is reflective.

You must ask:

“Did AI improve my thinking…
or replace my thinking?”

That difference matters.

Required Submission

You must submit:

1. PDF Presentation

Around:
10–15 slides/pages.

2. Workflow Diagram

Very important.

Clear.
Readable.
Logical.

3. Prompt Documentation

Screenshots.
Prompts.
Iterations.

4. Reflection Summary

300–500 words.

5. Update Your Google Site

Because your portfolio continues growing.

Important Reminder

“This assignment is NOT about making the most beautiful render.”

It is about:

• thinking,

• workflows,

• orchestration,

• reflection,

• decision-making.

CTA Bonus Layer

If you want bonus depth…

compare:

• ChatGPT vs Gemini,

• Gemini vs Claude,

• Midjourney vs DALL·E,

• different outputs,

• different interpretations.

Then ask:

Why are they different?

That is CTA thinking.

That is reflective orchestration.

Final Reminder

Slow down slightly here.

“The goal is not to show…”

how many AI tools you can use.

Pause.

“The goal…”

is to show:

how wisely
you can orchestrate them.

“Alright everyone…”

“Any questions before we end?” 

ARCH2360 — Assignment 3 Briefing Demonstration

CTA Comparison Presentation

+ Multi-Agent Reflective Analysis

Opening Briefing

“Alright everyone…”

“Now we move into Assignment 3.”

“And this assignment…”

is probably the most important intellectual assignment in this course.

Because now…

we move beyond:

• using AI,

• building workflows,

• generating outputs…

toward:

reflection,
comparison,
judgment,
and orchestration.

Assignment 3

CTA Comparison Presentation

• Multi-Agent Reflective Analysis

Weightage:
30%.

Group work:
4 students per group.

What Is This Assignment Really About?

“This assignment is not asking you…”

to find the “correct” AI answer.

Because in reality…

different AI systems may produce:

• different answers,

• different logic,

• different assumptions,

• different biases.

And that is exactly the point.

The Big Idea

“Most people use AI…”

to obtain answers.

Architects must learn to:

• question answers,

• compare perspectives,

• identify contradictions,

• and refine judgment.

That is what makes this assignment different.

Introducing CTA

This assignment introduces:

CTA
Cognitive Triangulation Architecture.

A framework where:

• multiple AI systems,

• multiple viewpoints,

• and human judgment

are combined into one reflective ecosystem.

Why Is This Important?

Because future architects…

may not work with:

• one software,

• one assistant,

• one system.

Instead…

you may work with:

• multiple AI agents,

• multiple consultants,

• multiple datasets,

• multiple simulations,

• multiple opinions.

And someone still must decide.

That someone…

is YOU.

Group Structure

Each group:
4 students.

Suggested roles:

Role

Responsibility

Student A

AI Exploration Lead

Student B

Comparative Analysis Lead

Student C

Visual / Workflow Lead

Student D

Reflection & Presentation Lead

But roles may overlap.

Because architecture itself is collaborative.

Your Task

Each group must select:

ONE design or built environment issue.

Examples:

• future housing

• flood resilience

• sustainable campus

• AI-assisted urban mobility

• smart public spaces

• construction automation

• adaptive reuse

• climate-responsive architecture

• future learning environments

Example Topics You Can Explore

Example 1 — Future Student Housing

Question:

“How should future student housing respond to AI-era lifestyles?”

Then compare:

• ChatGPT

• Gemini

• Claude

You may discover:

• one AI prioritizes efficiency,

• one emphasizes sustainability,

• one emphasizes psychology/community.

Interesting right?

Now YOU must evaluate.

Example 2 — Flood Resilience

Question:

“How can AI support flood-resilient urban systems in Klang Valley?”

Then compare:

• Perplexity

• Gemini

• ChatGPT

You may discover:

• different priorities,

• different assumptions,

• different urban strategies.

Again:
human judgment becomes necessary.

Example 3 — Smart Public Spaces

Question:

“Can AI-designed public spaces remain human-centered?”

Then compare:

• Midjourney

• DALL·E

• Firefly

One may generate:
beautiful but unrealistic spaces.

Another:
contextual but boring.

Another:
visually powerful but culturally disconnected.

Now the architect must think critically.

Minimum Requirement

You must compare:

MINIMUM 3 AI systems.

Examples:

• ChatGPT

• Gemini

• Claude

• Grok

• Perplexity

• Midjourney

• DALL·E

• Firefly

• Sora

What Must You Show?

1. Prompt Development

Show:

• original prompts,

• refinement prompts,

• contextual prompts.

And explain:

• why prompts evolved,

• how outputs changed,

• what improved.

Because prompting itself is a thinking process.

Example

Weak prompt:

“Design smart city.”

Too broad.

Refined prompt:

“Design a flood-resilient, pedestrian-friendly smart district for tropical Malaysian climate integrating passive cooling and public transport.”

Now the AI becomes more contextual.

2. Comparative Analysis

This is the heart of the assignment.

You must compare:

• similarities,

• contradictions,

• strengths,

• weaknesses,

• hallucinations,

• contextual differences.

Example Comparative Thinking

Criteria

AI A

AI B

AI C

Creativity

Strong

Medium

Strong

Buildability

Weak

Strong

Medium

Cultural Context

Weak

Strong

Medium

Sustainability

Medium

Strong

Medium

This comparison is important.

Because architects must learn:

not all AI outputs are equally meaningful.

3. Human Reflection & Judgment

MOST IMPORTANT SECTION

You must explain:

• What did the AI fail to understand?

• What required human judgment?

• What ethical concerns emerged?

• Which output became most meaningful?

• How did discussion improve the final outcome?

This is where your humanity enters.

Very Important

“Copy-paste is NOT authorship.”

Authorship requires:

• questioning,

• refinement,

• interpretation,

• responsibility.

Required Deliverables

A. Presentation Slides

15–20 slides recommended.

B. Comparative Matrix

Very important.

You must visually compare:

• AI systems,

• outputs,

• logic,

• strengths,

• weaknesses.

C. Reflection Summary

500–800 words.

This should sound:

• thoughtful,

• reflective,

• analytical,

• human.

Not robotic.

D. AI Prompt Appendix

Include:

• prompts,

• screenshots,

• workflow evidence.

E. A1 Poster + Google Site Update

Your semester portfolio continues growing.

This is why Assignment 1 was important.

Everything now becomes connected.

Presentation Structure

Suggested flow:

1. Introduction

2. Issue / Problem Statement

3. AI Systems Used

4. Prompt Strategy

5. Comparative Analysis

6. Reflection

7. Final Judgment

8. Conclusion

What I Want To See

I do NOT want to see:

• random AI images,

• random screenshots,

• blind acceptance.

I want to see:

• orchestration,

• reflective thinking,

• comparison,

• judgment,

• critical awareness.

Bonus Layer — CTA Depth

Groups may receive additional recognition if they demonstrate:

• deep triangulation,

• reflective dialogue,

• contextual architectural reasoning,

• ethical awareness,

• meaningful orchestration between AI systems.

This is where the assignment becomes powerful.

Final Reflection

Slow down slightly here.

“The future architect…”

may work with many intelligent systems.

Pause.

“But wisdom begins…”

when humans learn how to question them.

“Alright everyone…”

“Any questions?”