• Everyday AI
  • Posts
  • OpenAI's o3 Use Cases: How to use the world’s new most powerful LLM at your company

OpenAI's o3 Use Cases: How to use the world’s new most powerful LLM at your company

OpenAI's o3 use cases, Perplexity goes after Siri, Copilot AI updates, NVIDIA AI assistant plugins and more!

Sup y’all 👋

(though V long, so watch on 2x)

Had us thinking — should we do a comparison between Gemini 2.5 Pro and OpenAI’s new o3?

Or you not into that?

Should we do a comparison show between Google's Gemini 2.5 Pro and OpenAI's o3?

Login or Subscribe to participate in polls.

✌️

P.S. tomorrow’s show is with Pulitzer winning journalist Gary Rivlin, who wrote AI Valley. Don’t miss it!

Outsmart The Future

Today in Everyday AI
6 minute read

🎙 Daily Podcast Episode: We dive into the OpenAI’s o3, the world's most powerful LLM, showcasing use cases for your business. Give it a listen.

🕵️‍♂️ Fresh Finds: ChatGPT and Washington Post content partnership, Grok gets Vision and more. Read on for Fresh Finds.

🗞 Byte Sized Daily AI News: Perplexity goes after Siri with AI assistant, Microsoft Copilot gets AI overhaul and NVIDIA’s AI assistant gets plugins. For that and more, read on for Byte Sized News.

🧠 Learn & Leveraging AI: Want to take immediate advantage of OpenAI’s new o3 model? Here’s how you can use it to grow your business. Keep reading for that!

↩️ Don’t miss out: Did you miss our last newsletter? We talked about OpenAI’s o3 and o4-mini models, Google's new AI phase and Anthropic’s plan for AI employees. Check it here!

 OpenAI's o3 Use Cases: How to use the world’s new most powerful LLM at your company 💼

Wait. So.... how do you actually use OpenAI's new o3 model?

↳ It's legit agentic.

↳ Can think on its own.

↳ Use multiple tools in sequence.

This is not the normal blueprint for how a LLM works.

Don't worry -- we got you.

Also on the pod today:

• OpenAI's o3 Model Tiers Explained 💡
o3 vs. GPT 4 Benchmarking 📊
LLM Utility in Business 🗂

It’ll be worth your 1 hour and 14 minutes:

Listen on our site:

Click to listen

Subscribe and listen on your favorite podcast platform

Listen on:

Upcoming Episodes

Here’s our favorite AI finds from across the web:

New AI Tool Spotlight – Kiva is an AI SEO agent, Alter is an AI copilot for your Mac and Peek is an AI personal finance coach.

ChatGPT – OpenAI and The Washington Post have announced a content partnership to include Washington Post articles in ChatGPT responses.

xAI - Grok now has a Vision feature that pairs with Grok’s voice mode.

Perplexity – Perplexity has launched its iOS voice assistant. We tested it a bit. A bit wonky, but when it works…. super impressive.

If it gets better — this could squash Siri.

Apple – Apple’s new Siri chief Mike Rockwell is revamping the voice assistant’s management.

AMD – AMD will showcase new AI tech at Computex 2025.

Perplexity – Perplexity has released a new report on how it detects and counter malicious uses of Claude.

xAI - Grok now has a Vision feature that pairs with Grok’s voice mode.

Social Media – WhatsApp is defending its optional AI tool being added by Meta even though you can’t currently turn it off.

Trending in AI – Sam Altman is resigning as a chairman of the board of nuclear energy for Oklo.

1. Perplexity goes after Siri with smart AI assistant for iPhone

A week after teasing the idea, Perplexity AI has launched a new iPhone app version that performs many core Siri tasks—like managing Reminders, sending emails, controlling Apple Music, and even booking rides—without replacing Siri itself due to Apple’s system restrictions.

Unlike Siri, which is baked deeply into iOS and exclusive at the lock screen, Perplexity offers a fresh AI chatbot interface with real voice assistant capabilities that rival Siri’s usefulness. This move highlights growing innovation around Apple’s closed voice assistant ecosystem and hints at more practical AI tools arriving for iPhone users beyond Apple’s own offerings.

2. Microsoft 365 Copilot Gets a Major AI Overhaul 🧑‍🎨️️

Microsoft is rolling out a redesigned Microsoft 365 Copilot app that brings its business AI experience closer to the consumer version, featuring smarter chat memory, AI-powered enterprise search across multiple platforms, and a new Create feature using GPT-4o for generating images and more.

The app now includes Notebooks for project-based AI collaboration and an agent store simplifying access to AI tools, signaling a shift toward AI agents playing bigger roles in workplace productivity.

3. NVIDIA Expands G-Assist AI with Plugin Support 🔌

NVIDIA is rolling out a major update to its Windows-based G-Assist AI, transforming it from a gaming optimizer into a versatile assistant with plugin capabilities. The new ChatGPT-powered plugin builder allows developers to create custom tools for everything from Spotify control to Twitch streamer checks and real-time stock or weather info.

Running locally on RTX GPUs with 12GB VRAM minimum, G-Assist leverages small language models to keep data private and responsive.

4. Singapore Company Unveils First Truly Autonomous AI Agents 🤖

A Singapore-based company, Butterfly Effect, has unveiled Manus, one of the first truly autonomous AI agents that can research, decide, and act with minimal human input. While this marks a leap in AI capabilities, the technology raises urgent privacy and security concerns, especially as personal data may already be scraped and used without consent—posing risks like data leaks and prompt injection attacks.

Experts warn that Manus’s autonomy amplifies these dangers, making it harder to detect misuse or breaches, with troubling links to servers in China adding geopolitical worry.

5. Google’s Gemini Hits 350M Monthly Users 📈

Google’s AI chatbot Gemini has surged to 350 million monthly active users as of March 2025, according to internal figures revealed in an ongoing antitrust case, reported by The Information. This rapid growth—from 9 million daily users in October 2024 to 35 million just last month—shows how deeply Gemini is integrating into everyday tech.

Despite this expansion, it still trails industry leader ChatGPT, which boasted around 600 million monthly users in the same period.

6. MIT Unveils Periodic Table for Machine Learning Algorithms 🧠

A new study from MIT researchers introduces a groundbreaking periodic table that maps over 20 classical machine-learning algorithms by revealing a shared core equation, offering a fresh way to understand and innovate AI models. This framework doesn’t just organize existing methods; it predicts undiscovered algorithms and has already led to an image-classification algorithm outperforming current state-of-the-art techniques by 8%.

By simplifying how algorithms relate data points, this discovery could accelerate AI development, helping researchers and companies build smarter, more efficient models without reinventing the wheel.

🦾How You Can Leverage:

The future of work just got WEIRD. 

After finally getting a dozen or so hours of test time in on OpenAI’s new o3 model, we were ready to dive in and give you LIVE business use-cases

This thing’s freaking agentic. A straight up silly piece of technology that doesn’t make sense. 

And like we talked about yesterday in Part 1 of our two-part series, OpenAI’s new o3 model is the most powerful AI model in the world. 

OpenAI's new o3 model isn't just smarter than previous AI—it's truly agentic, switching between tools like web search, Python, and image generation all by itself.

In today's (kinda rambling) Everyday AI episode, we put this beast through increasingly ridiculous tests to find out if it's actually the category-changing model everyone claims. (Watch on 2X speed. We won’t be mad.) 

 Spoiler alert: o3 is even better than you think an LLM could be. 

Let’s goooooooooo👇

1 – Meet the AI That Uses Multiple Tools like a Handyman ⚒️

Unlike old-school transformer models, o3 doesn't just complete your sentences—it REASONS with multiple tools on demand.

Step by step. Tool by tool. Like your smartest coworker on espresso who’s actually/literally/physically juggling physical objects.

When we fed it complex tasks, it seamlessly jumped between tools, running Python code, searching the web, going back to data viz with Python, back to reasoning, back to searching the web…. and creating visualizations without being told what to do next.

WTF.

The o-series models aren't just smarter. They're fundamentally different beasts. Don’t get us wrong — Google’s Gemini 2.5 is a very capable model that can also use tools agenticly.

The biggest difference?

Google’s Gemini 2.5 Pro uses tools sequentially and only once. OpenAI’s goes back and forth and back and forth and back and forth.

(That makes a YUUUUUUUUGE difference.)

On paid ChatGPT Plus plans, you get 50 weekly messages with this digital sorcery, though. (Though we heard rumblings that mighta been upped today.)

Pro users ($200/month) get nearly unlimited access.

Try This

Feed o3 your messiest problem—the kind requiring multiple tools to solve. Try: "Analyze our customer retention data, identify what's causing churn, and create a visual dashboard showing high-risk segments."

Then sit back and watch the magic happen.

2 – Three “That’s Impossible” Tasks that o3 Crushed 💪

We uploaded a complex PDF filled with images, charts, and tiny logos. o3 transcribed it flawlessly—every data point, every design element.

Then we went bigger. "Find our 20 latest podcast episodes, summarize them, and identify trends."

It delivered complete episode details AND five cross-episode patterns that would have taken hours to spot manually.

But the real showstopper? We uploaded podcast stats from 23,000 cities. Hundreds of thousands of data points.

Within minutes, o3 gobbled all that data and revealed that Friday episodes get 9% more plays, "AI agents" keywords boost performance by 50%, and Australia is our surprise #2 market—then casually built an interactive dashboard we could actually use. 

O3 really went down under for these tricks. lolz

Try This

Upload your messiest dataset (CSV works perfectly). 

Ask for "3 obvious trends, 3 non-obvious trends, and 3 growth opportunities." The magic happens when you ask for "non-obvious" insights—that's when o3 thinks laterally about your business in ways you haven't considered.

3 – The Day AI Wrote Our Newsletter for Us 🥹

Welp that was awkward.

We showed o3 three examples of our newsletter format and a Boolean search URL for finding AI news.

"Write today's newsletter."

What happened next was borderline creepy. And had us scratching our heads like WAIT WTF. 

  • Perfect formatting.

  • Accurate news summaries.

  • Proper emoji usage.

  • All featuring ACTUAL breaking news from today.

  • Plus a working interactive web version with search filters that actually functioned.

A task that takes our team hours—done in seconds.

(You comin for our jobs CHAD!?!)

Try This

Create a Boolean search string for your industry news (like "Microsoft OR Apple OR Google AND AI" limited to 24hrs).

Feed it to o3 with examples of your communication style. Ask it to create your daily industry briefing with highlighted competitive threats specific to YOUR business. Instant personalized intelligence.

This isn't just incremental progress—it's a quantum leap. Seismic step.

The future isn't about who can do the work.

It's about who knows what to ask for.

Reply

or to participate.