1 of 35

Navigating ChatGPT Capabilities and User Interface

2 of 35

What We'll Cover Today

Know the tools

The chat interface and what each built-in tool does.

Prompt with purpose

Be specific, give context, ask for the format you want.

Put it to work

Summarise, analyse data, read images, files and video.

You will leave having done all three.

3 of 35

Prompt Input Tools

Type, attach a file, or switch on a tool before you ask.

Conversation Management

Every chat is saved and searchable — one chat per topic.

4 of 35

Edit, Don't Restart

Reword or regenerate — you keep the context.

5 of 35

Deep Research - Produces structured, multi-source research reports with clear insights

Shopping Research -Compares products, prices, and reviews to guide purchases�
Web Search -Finds and summarises real-time information from the web�
Study & Learn - Explains concepts step-by-step like a personal tutor�
Canvas - A focused workspace to draft and refine long content�
Quizzes - Creates interactive quizzes to test understanding quickly�
Explore Apps - Access specialised AI tools built for specific tasks�
Create Image - Generates visuals from text prompts for ideas and design�
Add Photos & Files - Upload files or images for analysis and explanation�
Thinking - Reasons through complex problems before giving structured answers

6 of 35

7 of 35

8 of 35

9 of 35

1. Memory

ChatGPT can remember useful preferences or ongoing context to improve future responses.

Examples:

Preferred tone (formal, friendly, concise)
Repeated interests or projects
How detailed you like explanations�

You’re always in control — memories can be viewed, edited, or deleted.

2. Custom Instructions

You tell ChatGPT how you want it to respond and what it should know about you.

Examples:

“Explain things simply, no jargon”
“Write in a professional, Singapore-context tone”
“Assume I’m an adult learner, not a beginner”�

This applies across all chats.

3. Response Style & Personality

Adjusts how ChatGPT communicates — not what it knows, but how it speaks.

Examples:

More concise vs more detailed
More structured vs conversational
Teaching-style vs consultant-style�

10 of 35

Beginner Prompting with ChatGPT

11 of 35

Why Prompting Matters

ChatGPT works best when instructions are clear
It does not guess what you want
Better prompts = better answers
Clear prompts save time and reduce frustration

12 of 35

Common Beginner Challenges

“The answer is not what I want”
“It’s too complicated”
“I don’t know how to ask properly”

You don’t need technical skills — just a simple structure

13 of 35

Prompting ChatGPT is like talking to an expert or consultant.

�You need to explain your situation clearly, provide enough context, and state the outcome you want.

�Vague questions give vague answers.

Clear requests give better results.

14 of 35

Checkpoint: Before We Practise

Rewrite one prompt

Add role, context and format to a question you asked this week.

Swap with a neighbour

Would they know exactly what you want? If not, sharpen it.

Name your test case

One real task you will bring to the hands-on activities.

Vague in, vague out — clear in, useful out.

15 of 35

Exploring GPT

Play Prompt

16 of 35

Write a Short Story

“Write a short 150-word story about an ice cream seller on Orchard Road. Use simple English.”

17 of 35

Languages, Style & Tone

Translate the story in (Mandarin, Tamil, Malay, French, etc).

Rewrite the story in (Singlish / Formal English / Use expressive shakespearian style / legal language)

Rewrite to make it a humorously or sad story

18 of 35

Response to Correspondence

Subject: Complaint about poor service during my stay at Good Service Hotel

Dear Manager,

I am writing to express my disappointment with the poor service I received during my recent stay at the Good Service Hotel. I stayed at your hotel from March 1st to March 5th and unfortunately, my experience was marred by the poor service provided by one of your staff, Alice. Throughout my stay, Alice was consistently unhelpful and unfriendly. On the first day, I asked her for directions to a nearby restaurant and she seemed annoyed and unwilling to provide assistance. On another occasion, I requested a late check-out and she rudely informed me that it was not possible, despite the fact that I had seen other guests being granted late check-outs. As a customer who has paid for quality service, I am extremely disappointed with the poor service provided by Alice during my stay at your hotel. I believe that a hotel of your reputation should have staff who are friendly, helpful, and attentive to the needs of guests. I would like to request that you take appropriate action to ensure that such poor service is not provided to other guests in the future. I hope to receive a response from you soon regarding the steps that you will be taking to address this matter.

Sincerely,

Adam

You are the Manager of Good Service Hotel. You received the following email from your customer. After investigation, you did not find Alice did anything wrong, however you also want to win back the customer. Write an appropriate reply to the customer.

Share your responds

19 of 35

Current Affairs

Prompt:

Summarize today’s top 3 news headlines in Singapore

Share with Partners

Are they the same?

Can they be trusted?

20 of 35

The Power of Summarization in GPT

What is Summarization?

Condensing lengthy information into brief, clear content while keeping the main meaning.

What ChatGPT Can Do:

Summarize texts and highlight key points for different audiences.

Why It’s Useful:

Saves time, reduces overload, and enhances understanding.

21 of 35

Common Use Cases for Summarization

Work & Office

Summarize long emails or meeting notes
Create executive summaries

Learning & Training

Summarize articles or study materials
Turn complex topics into beginner-friendly notes

Daily Life

Summarize news articles
Shorten terms & conditions

Business

Summarize customer feedback
Review policies or reports quickly

22 of 35

Inputs

FILES – EXCEL, CSV, DOCX, PPTX, PDF, ETC

IMAGES –

URL

YOUTUBE

23 of 35

Hands-on Practice: Try It Yourself

Try these prompts:

“Summarise this in 5 bullet points.”

“Explain this in simple language for a beginner.”

“Give me a one-paragraph summary.”

“What are the key takeaways?”

Compare the original with the summary — how much time did you just save?

24 of 35

A Concise History of Artificial Intelligence

Artificial Intelligence (AI) refers to the scientific and engineering pursuit of creating machines capable of performing tasks that typically require human intelligence. These tasks include reasoning, learning, perception, language understanding, and decision‑making. The history of AI is not a linear march of progress but a cyclical story of optimism, technical breakthroughs, limitations, and renewed paradigms. Understanding this history provides critical context for today’s rapid advances and ongoing debates surrounding AI’s capabilities, risks, and societal impact.

Sample text for the summarisation activity — do not read this aloud.

Tell participants to copy the passage or use the file you shared, then try: “Summarise this in 5 bullets”, “Explain it to a 12-year-old”, and “Build a timeline of the key milestones”.

Good follow-up question for the room: which facts did the summary drop, and did that matter?

Full sample passage (share as a handout or paste into chat):

A Concise History of Artificial Intelligence

Introduction

Artificial Intelligence (AI) refers to the scientific and engineering pursuit of creating machines capable of performing tasks that typically require human intelligence. These tasks include reasoning, learning, perception, language understanding, and decision‑making. The history of AI is not a linear march of progress but a cyclical story of optimism, technical breakthroughs, limitations, and renewed paradigms. Understanding this history provides critical context for today’s rapid advances and ongoing debates surrounding AI’s capabilities, risks, and societal impact.

Intellectual Origins (1940s–1950s)

The conceptual foundations of AI emerged in the mid‑20th century at the intersection of mathematics, logic, neuroscience, and early computing. Alan Turing’s seminal 1950 paper, Computing Machinery and Intelligence, posed the provocative question, “Can machines think?” and introduced the Imitation Game (now known as the Turing Test) as a behavioral criterion for machine intelligence.

Early computational models such as McCulloch and Pitts’ artificial neurons (1943) and Norbert Wiener’s work on cybernetics framed intelligence as an information‑processing phenomenon. These ideas suggested that cognition might be mechanized, a radical notion at the time.

The Birth of AI as a Field (1956–1960s)

The term “Artificial Intelligence” was formally coined in 1956 at the Dartmouth Summer Research Project on Artificial Intelligence, organized by John McCarthy, Marvin Minsky, Claude Shannon, and Nathan Rochester. The participants believed that significant aspects of intelligence could be precisely described and implemented in machines—a belief that fueled early optimism.

During this period, symbolic AI dominated. Researchers focused on rule‑based systems that manipulated symbols to emulate human reasoning. Early successes included programs for theorem proving, game playing (notably checkers and chess), and basic natural language processing. Systems such as ELIZA demonstrated that even simple pattern‑matching could produce surprisingly human‑like interactions.

Expansion and First AI Winter (1970s)

Despite early promise, symbolic systems struggled with real‑world complexity. They required extensive hand‑coded rules, were brittle outside narrow domains, and lacked common‑sense reasoning. As expectations outpaced results, funding agencies grew skeptical.

The 1970s marked the first “AI winter,” a period of reduced investment and enthusiasm. Reports such as the UK’s Lighthill Report (1973) criticized AI’s limited practical impact, leading to funding cuts and a contraction of the field.

Expert Systems and Commercial Revival (1980s)

AI regained momentum in the late 1970s and 1980s through expert systems—rule‑based programs designed to emulate the decision‑making of human specialists. Systems like MYCIN (medical diagnosis) and XCON (computer configuration) demonstrated tangible commercial value.

However, expert systems were expensive to build and maintain, highly domain‑specific, and difficult to scale. As hardware costs fell and maintenance burdens rose, interest again waned by the late 1980s, triggering a second AI winter.

The Statistical and Data‑Driven Turn (1990s–2000s)

A significant shift occurred as AI increasingly embraced statistical methods and machine learning. Rather than relying on handcrafted rules, systems began learning patterns from data. Advances in probability theory, optimization, and computational power enabled techniques such as decision trees, support vector machines, and Bayesian networks.

This era saw notable successes in speech recognition, computer vision, and information retrieval. IBM’s Deep Blue defeating world chess champion Garry Kasparov in 1997 symbolized the power of specialized, data‑intensive AI, though such systems remained narrow in scope.

Deep Learning and the Modern AI Renaissance (2010s)

The modern resurgence of AI is closely tied to deep learning—neural networks with many layers trained on large datasets using powerful GPUs. Breakthroughs in image recognition, speech processing, and natural language understanding followed landmark results such as the 2012 ImageNet competition.

Systems like AlphaGo, which defeated the world’s top Go players, and large language models capable of generating coherent text marked a qualitative leap in performance. Unlike earlier systems, these models exhibited emergent capabilities not explicitly programmed, reigniting discussions about general intelligence.

Contemporary Challenges and Reflections

Today’s AI systems are widely deployed across science, industry, and society. Yet their history cautions against uncritical optimism. Persistent challenges include interpretability, bias, robustness, energy consumption, and alignment with human values. While current models excel at pattern recognition and generation, they still lack genuine understanding, intentionality, and moral agency.

Conclusion

The history of AI is a story of evolving ideas about intelligence itself—from symbolic reasoning to statistical learning and large‑scale neural computation. Each phase contributed tools, insights, and cautionary lessons. For scholars and practitioners alike, appreciating this historical trajectory is essential to responsibly shaping AI’s future, balancing innovation with realism, and grounding technological ambition in intellectual humility.

25 of 35

ChatGPT’s Ability to Analyse Data & Reports

What Does “Analyse” Mean?

Read information carefully
Find patterns, trends, and problems
Explain numbers in simple language
Highlight what matters for decision-making

What ChatGPT Can Analyse:

Financial reports (profit, loss, expenses)
Sales data (daily, monthly, yearly)
Tables, spreadsheets, and charts
Business and performance reports

Why This Is Powerful:

Saves time reading reports
Makes numbers easier to understand
Helps people make faster decisions

26 of 35

�Hands-on Practice: Analyse Sales Data with ChatGPT

Try these prompts:

“Summarise the overall sales performance.”

“Which product sold the most and the least?”

“Which category generated the highest revenue?”

“What sales trends do you notice?”

Now push further: “Show this as a chart.” “What should we do next quarter?”

“

27 of 35

Date	Product Name	Category	Quantity Sold	Unit Price ($)	Total Sales ($)
01-Jan-25	Coffee Mug	Home	25	8	200
01-Jan-25	Water Bottle	Home	18	12	216
02-Jan-25	T-Shirt	Apparel	30	15	450
02-Jan-25	Notebook	Stationery	40	5	200
03-Jan-25	Coffee Mug	Home	20	8	160
03-Jan-25	Backpack	Accessories	10	45	450
04-Jan-25	T-Shirt	Apparel	22	15	330
04-Jan-25	Water Bottle	Home	15	12	180
05-Jan-25	Notebook	Stationery	50	5	250
05-Jan-25	Backpack	Accessories	8	45	360

Sample Sales Data

28 of 35

Hands-on Practice: Analyze

Perform a comprehensive analysis of (company Name)’s FY20XX Annual Report. In your report include

Executive summary: key highlights and overall performance
Financial Reviews: trends, profits, with year on year comparisions
Strategic initiatives: major investments, M&A, Digital transformations
SWOT assessments
Management outlook
Valuation and Recommendations.

The report to be written as a email report to an analysis team and placing key figures into a table format for easy reading.

Source:

https://www.dbs.com/iwov-resources/images/investors/annual-report/dbs-annual-report-2024.pdf?pid=splitter-home-annual-report-2024

29 of 35

Other Hands-on Practice

Generate Ideas or Brainstorm

a. List 10 free weekend activities for a family in Singapore

b. This is what’s in my fridge – what recipe do you have?

Trip Planning – Create Itinerary

Help me plan a 3 day trip to Malacca with food and shopping included.

Teach Me / Explain

Explain to me what is inflation in simple terms. Give me examples.

Role Play

Pretend to be a shopkeeper. Role play a customer asking for discount.

30 of 35

Cool Stuff

31 of 35

Put the data in a table format. Analyse the data and provide a summary for management.

Create visual charts and dashboard slide

32 of 35

compile the receipts into a summarize table format

33 of 35

Dear All,

Please find below the meeting minutes from our discussion on planning the New Year Party.

Meeting Details

Date: 23 December 2025
Location: Tampines Office

Attendance

Andrew
Tom
Jill

Agenda

Planning for New Year Party

Discussion & Agreements

The party will be held at Tom’s home.
The event will be organised as a potluck party.
Family and friends will be invited to join the celebration.

Action Items

Jill to compile the list of attendees and confirm the total number.
Andrew to create a WhatsApp chat group for coordination.
Tom to assign food and items for attendees to bring to the party.

Please let everyone know if there are any corrections or additions. Looking forward to making this a fun and memorable celebration.

Best regards,�Andrew

34 of 35

https://www.youtube.com/watch?v=fMcekic_V8k

Summarize this Episode

35 of 35

Three Habits to Take Away

Be specific

Say exactly what you want — length, audience, format.

Give context

Brief it like a new colleague, not a search engine.

Always verify

Check any name, number or date before you use it.

So — what is the one thing you will try this week?