Issue 3: AI Automation and Effortless Presentations
Welcome back to our third issue! AI UnGeeked has been a wild ride so far, wilder than we imagined. Every time we start drafting a new issue, we think “gonna keep it short and simple”, and then like any best laid plan, we figure otherwise.
Anyway, there’s lots in store for today, so let’s start off!
News That Matters
Personalized product recommendations with images, reviews, and direct purchase links - Sounds like Google? It’s actually Open AI. And there comes the most direct challenge yet to Google search and its advertisement revenues.
The newly launched Meta AI app is positioned as an assistant with ability to personalise and contextualise support for users. It also comes with a Discover feed where users can share their AI journeys with others. Meta marrying what it does best, social media, with what it wants to beat everyone else at, AI. But it does not have access to the web or real-time information, yet. Interesting times indeed.
“I am so proud of you, and I honour your journey,” ChatGPT apparently said in response to a user updating that they stopped taking medication. A number of users similarly highlighted ChatGPT’s latest update to GPT-4o as being overly flattering and ‘sychophant-y’. Open AI has now rolled back this update.
One Trend of Note
No-code, Low-code AI: Automation for everyone?
Ever thought "I wish I could automate this boring task”? But you’re not a tech person? Good news: no-code and low-code automation tools, now powered by AI, are here to help.
These tools let you automate repetitive tasks with little to no coding — using drag-and-drop steps, plain language prompts, or ready-made templates. You can connect your everyday apps (like Gmail, Slack, and Google Sheets) and set simple instructions so things just happen. This isn’t about building complex systems — it’s about automating the micro-stuff: copying updates, chasing replies, filling out the same docs again and again.
So what does that look like in real life?
A typical automation might look like this:
→ pull data from one or multiple sources (spreadsheets, email, forms etc.)
→ use AI to process it in some way (summarize or analyze or extract insights)
→ share the output (via email, messaging app or even a chatbot)
Here are a few practical examples:
Pull data from Google Sheets weekly, use Perplexity (or your favourite AI model) to write a summary, and email it to your team — no manual copy-pasting and summarising needed.
Collate emails (from specific people or teams), summarise into a daily digest, send to a Slack channel every morning.
Scan incoming vendor invoices, extract details like due dates and amounts, and log them in a tracker. Flag anything odd — like duplicates or missing info.
But wait, don’t we already use automation?
Yes, but most of it is rule-based. Think: “If a lead comes in, send a welcome email.” It’s great for predictable tasks, but that’s where it ends. AI-powered automation goes a step further, and can handle tasks that aren’t so cut & dry.
For example:
It can read a support email, detect urgency, and assign it to the right team.
It can scan an invoice and file it under the right category — even if it looks a little different than usual.
In short: Rule-based is rigid. AI automation is smarter and adapts to context.
So what are these magical no-code, low-code apps?
If you’re just getting started, tools like Zapier and Make are easy to use and free to try.
If you want a bit more control and customization, Microsoft Power Automate and n8n offer more flexibility, especially if you are somewhat familiar with coding.
Bottom line
AI + no-code/low-code tools are changing how work gets done — putting automation in the hands of regular teams, not just IT. It’s a real shift, and it’s gaining momentum.
Curious to dive deeper? Stay tuned. We plan to cover how to actually use this in your everyday work in future issues.
AI Term of the Fortnight
What’s Behind All This AI Magic? Meet Transformer Models
If you’ve asked an AI tool to write a poem, or used a tool that turns your scribbles into a polished image, you’ve experienced the power of something called a Transformer model — the tech behind most of today’s Generative AI.
Let’s break it down.
Generative AI (Gen AI for short) means AI that can create things — like text, images, code, music, and more. Not just analyze data, but actually make something new. It’s why you can ask an AI to draft a summary or come up with five tagline options, and it does it instantly. The tech that makes this possible? Transformer models.
Unlike older models that struggled with long texts or complex patterns in texts, Transformers excel at understanding context within text — figuring out which words in a sentence matter most or how different ideas connect. It's kind of like how our brain zooms in on key phrases while reading quickly. That’s why Transformers are perfect for tasks like chatbots, translation tools, and more.
Then there’s GPT — short for Generative Pretrained Transformer. It’s been trained on a large amount of text — books, websites, articles — so when you ask it a question, it doesn’t look up an answer. It generates one on the spot, based on patterns it has learned. The big names in AI — OpenAI’s GPT models, Google’s Gemini, X’s Grok, Anthropic’s Claude — are all based on this GPT structure: they’re generative (they create content), pretrained (trained on huge datasets), and Transformer-based.
In short, Transformer models are the foundation of today’s AI superpowers. They’re complex on the inside but are becoming easier and more intuitive for us to use, without requiring technical knowledge.
AI in Practice
Gamma.App: From Draft to Deck in Minutes
A powerful way to engage AI is to combine the strengths of many of these apps. For instance, use ChatGPT or Gemini to draft the structure and content for a presentation, and then employ Gamma.App to create a presentation using that content.
Here’s how:
Select a topic you know well. Remember — AI tools are for support, not a replacement for thinking.
Write the prompt clearly specifying a list of points you want covered.
Feed this prompt to an AI app of your choice e.g., ChatGPT or Gemini.
Copy the output and feed into the Gamma.app text editor. Tweak the structure / text if needed.
Select your slide count, density of text per slide and whether to add contextual images (from the internet or AI generated ones)
Pick a presentation theme based on your audience and the aesthetic they’d prefer. This sets the presentation’s colour scheme and slide background.
Click generate, and voila you have your first draft! You can now edit the presentation easily — using convenient drag and drop widgets that auto align, without you having to spend precious minutes moving each box to the ‘n’th millimetre.
Share the final presentation to recipients’ email ids, or present directly to the audience.
This combo is perfect for business professionals, consultants, and educators who need quick, high-quality presentations.
Try it: ChatGPT or Gemini + Gamma.app
And if you want a deeper dive on using Gamma.app, we have a tutorial here (replete with screenshots and detailed steps).
With that, we wrap up Issue 3 of AI UnGeeked. As always, tell us what has worked so far, what hasn't, and everything in between. And if you have topics you want us to touch upon, we are just a comment away!
Cheers!