CODED × Alshaya · Kuwait
CODED×ALSHAYA

From Blank Page
to Finished Deck

Day 2 — Presentations, data analysis, and the prompt library you keep using after the workshop ends.

DAY 2 Wednesday 2 September 2026 09:00 – 13:00 CODED Campus
Ali Taqi · Alshaya & Al-Injaz
press → to begin · ? for controls

Callback: yesterday's CTFT prompt

You wrote Context, Task, Format and Tone for one real document. Today you reuse that exact structure all day — to build a presentation, to question a set of data, and to screen a shortlist.
Same framework. New outputs.
Day 2

A full presentation used to take two to three hours.

Today you'll build one before the first break.

01
AI for Presentations

The presentation problem

Blank slides, no structure, no design skill, and no time. Most people spend two to three hours building one deck — and most of that goes on layout, not on thinking.
That time comes straight out of your real workday.
The landscape

Four ways to let AI build the deck

We will use the first two today, so you can feel the difference between them.

Claude Design
Anthropic
Describe the deck inside the chat you are already working in. You get an editable visual canvas — click any element, change it, export it. Your context is already there.
We use this
Gamma
Gamma.app
A tool built only for presentations. Type a prompt, pick a theme, get a designed deck — then export it to PowerPoint or PDF.
We use this
Beautiful.ai
Beautiful.ai
Template-led. The layout adjusts itself as you add content, so slides stay tidy without you nudging boxes around.
Canva AI
Canva
Content and design in one place, with a very large library of images and brand assets behind it.
Different tools, one skill. The prompt is what carries across.

The hybrid approach

Use AI to generate the outline and the first draft. Then add your own expertise, your own numbers, and your own judgment before anyone else sees it.
AI drafts. You own the version that gets presented.
🎨
Live demo

Live demo — Claude Design

Same chat, same context. Describe the deck and watch it become a canvas you can actually edit.

◆
It already knows what we have been discussing — no re-explaining the situation
◆
Click any element to change it, rather than regenerating the whole slide
◆
Watch how a vague instruction produces a vague layout — the Day 1 lesson, again
⚡
Live demo

Live demo — Gamma

A tool that does one thing. Prompt in, themed deck out, ready to export.

◆
One prompt produces a full deck with a theme applied, in well under a minute
◆
Swapping the theme restyles every slide at once
◆
Export to PowerPoint — and notice what needs fixing after the export
Framework

What goes into a presentation prompt

Type and length — “a 5-slide presentation”
Topic and key points — the specific ground it must cover
Audience and purpose — who is in the room, and what you need from them
Tone — professional, persuasive, plain
Compare

Weak prompt vs. strong prompt

Weak
“Make a presentation about maintenance.”
Strong
“Create a 5-slide presentation for property managers about the new maintenance request system. Cover the problem, the new process, staff responsibilities, and the launch date. Use a professional tone.”
That is CTFT again, wearing a different hat. Context, task, format, tone.
Lab

Lab D2.1 — Report to Deck

15 minutes · in the Day 2 lab

1
Attach yesterday's quarterly report to a new Claude chat
Download it from the lab — the same five-page PDF you verified yesterday.
2
Write your own CTFT prompt, then ask Claude to design it
A six-slide management summary for the operations director.
3
Open two slides and check every figure against the PDF
A deck that looks right but misquotes a number is worse than no deck.
4
Fix one slide by hand
So you feel where Claude stops and you start.
Open lab task D2.1 ↗
Lab

Lab D2.2 — Same Prompt, Different Tool

10 minutes · in the Day 2 lab

1
Copy the exact prompt you wrote in D2.1
The point is the prompt, not the product.
2
Generate the deck in Gamma
Pick a theme, let it build.
3
Put the two decks side by side
Which produced the better first draft? Which needed less fixing?
4
Export the Gamma deck to PowerPoint
See what shifts — spacing, fonts, charts.
Open lab task D2.2 ↗
Break
Stretch and Reset
15:00
15 minutes — back on time, please
02
Presentation Review & Advanced Features
Group discussion
What surprised you about the deck AI built for you?
one strength and one limitation from each table
Choosing

Which one, when?

🎨
Claude Design
You are already deep in the content with Claude, and you want to shape the deck visually without leaving the conversation.
⚡
Gamma
You want a themed, presentable deck fast — and you need it out as PowerPoint or PDF at the end.
📊
PowerPoint
Company template, exact control over every element, or highly technical detail that must sit exactly where you put it.
The honest rule: AI for the first draft, your own tool for the final one — whenever the deck really matters.

Features worth knowing

📥
Import
Paste in an existing outline or document and let it structure the slides for you.
✏️
AI rewrite
Select one slide and ask for shorter, simpler, or more detail — without touching the rest.
🎭
Themes
Restyle the whole deck with one click when the first look is not right.
📤
Export
Out to PowerPoint or PDF. Always check the export — spacing and fonts can shift.
🔒

Before you put real company data in a deck

CautionDo not paste tenant, financial or contract data into a public AI tool. Build the structure with sample data first, then add the real numbers yourself, afterwards, in your own file.
Same rule as Day 1. A new tool does not change it.
03
Data Analysis With AI

The data analysis problem

A spreadsheet with 200 rows lands in your inbox. No time to clean it, chart it, or work out what it means before the meeting starts.
Most people spend more time preparing data than analysing it.
Compare

What AI can and cannot do with data

AI cannot
Know whether a result is good or bad for your business, understand your organisation's context, or fix data that was wrong before it arrived
AI can
Calculate statistics, clean and sort, spot patterns and outliers, build charts, and draft the explanation
It will also do arithmetic confidently and occasionally get it wrong. Spot-check every number that matters.
Framework

Three rules for analysing data with Claude

One clear question at a time — not five at once
Spot-check every calculation — by hand, on the number that matters most
Ask it to show its working — “list every figure you used and where it came from”
Lab

Lab D2.3 — Finance: Monthly Performance

20 minutes · in the Day 2 lab

1
Download the workbook from the lab, upload it to Claude
Six sheets of sample data — safe to upload. Start with Occupancy & Income.
2
Work the three steps — one CTFT prompt each
Find the trend, rank the budget variances, draft the management summary.
3
Verify against the sheet
Recalculate at least one variance by hand before you trust it.
4
Record what you found in the lab
If a bullet has no number behind it, it does not go in.
Open lab task D2.3 ↗
04
Facilities Data & the AI + Excel Hybrid

Why hybrid beats either tool alone

Claude explains what a pattern means and writes it up in plain language. Excel proves the number is actually correct. Use both, in that order.
Trust, but verify. Every time, no exceptions.
Lab

Lab D2.4 — Facilities: Logs to Report

20 minutes · in the Day 2 lab

1
Point Claude at the Maintenance Log sheet
68 raw tickets in the same workbook — no cleaning, no pivot tables.
2
Write the weekly management report
Open tickets by category, anything ageing, contractor performance, what needs a decision.
3
Verify the counts yourself
Against the sheet, before you trust the write-up.
4
Draft the tenant notice about the lift work
Under 120 words. Clear and calm, apologise once — not three times.
Open lab task D2.4 ↗
Break
Stretch and Reset
10:00
10 minutes — back on time, please
05
Prompt Libraries & Your Department

Why save your best prompts?

A prompt that worked once will work again. Save it, and next time you skip straight past the rewriting to the result.
This is your prompt library — you started it on Day 1 without noticing.
Framework

What makes a saved prompt worth keeping

A title you will recognise later — “monthly maintenance update”, not “prompt 3”
The full CTFT prompt, with placeholders where the details change
A note on which tool it worked best in — and what you had to fix afterwards

Where this lands in your job

👥
HR
Onboarding letters, policy summaries, job description drafts, training announcements.
💰
Finance
Variance commentary, budget summaries, turning a table into a paragraph a board can read.
🔧
Operations
Maintenance updates, contractor briefs, schedules, incident write-ups.
📣
Communications
Tenant notices, announcements, FAQs, and the same message rewritten for three audiences.
Every one of these is draft, rewrite, or summarize — the three moves from yesterday.
Lab

Lab D2.5 — HR: Screening Support

20 minutes · in the Day 2 lab

1
Point Claude at the Candidates sheet
12 candidates in the workbook — IDs only, deliberately no names.
2
Build a reusable candidate-summary template
With [PLACEHOLDERS], so it works for any candidate and any role.
3
Screen one role against clear criteria
Meets / partly meets / does not meet — with one line of reasoning each.
4
Draft eight interview questions
The hiring decision never leaves the hiring manager.
Open lab task D2.5 ↗
Lab

Lab D2.6 — Build Your Prompt Library

15 minutes · in the Day 2 lab

1
Collect the prompts that worked best today
The deck, the finance summary, the facilities report, the screening template.
2
Save four things for each
A title, the full CTFT prompt with [PLACEHOLDERS], the best tool, and what you fixed.
3
Add one prompt for your own department
Something you will genuinely reuse next week.
4
Save it where you will find it
A document, a notes app, a shared drive — anywhere but your memory.
Open lab task D2.6 ↗
Today you built:

Three things you can use on Sunday.

A presentation generated in two tools and reviewed properly
Three department use cases — Finance, Facilities and HR — verified against the workbook
The start of a reusable prompt library, built around your own job
Tomorrow: put all three days together into one implementation package.
End of Day 2

See you tomorrow.

Day 3 — The Implementation Capstone

Thursday 3 September 2026 · 09:00 – 13:00
‹ Day 2
Exercise2:00
1 / 0
→ next ← back F full ? help
screen blanked — press B to resume

Presenter & pointer controls

→SpacePg Dn
Next slide
←Pg Up
Back
L
Laser pointer (follows the mouse)
B.
Blank the screen
F
Fullscreen
S
Sound on / off (starts off)
T
Start the exercise timer (task length on a lab slide, otherwise 2:00)
?
This panel
Esc
Close / resume
Most USB presenter remotes send Page Up / Page Down — they already drive this deck.
☕ Start a break
Pick a length — the timer counts down and shows everyone the return time.
Break
Break
10:00
Back at —