- The Automated
- Posts
- Meet ChatGPT’s new ‘Deep Research’ agent!
Meet ChatGPT’s new ‘Deep Research’ agent!
Plus, unlock how we built The Automated agency with AI + 50% discount for our members

Hello and welcome to the Automated, your AI tour guide.
After a stormy week, thanks to China’s DeepSeek nearly raining on OpenAI’s parade, the company has made a strong comeback with the perfect response.
First, they launched the new o3-mini reasoning model for all users, including free-tier users, and wait for it… they also unveiled a brand-new AI Agent that can actually do deep, really deep research for you.
Here’s what we have for you today:
💃 OpenAI’s latest AI Agent that actually does its homework!
🥳 OpenAI rolls out O3-Mini, a new reasoning model with a free ChatGPT version.
🦾 How to build AI Agents for free.
📈 How to use Claude Artifacts to visualize data.
🤖 ChatGPT Prompt Of The Day: Validate Data Format.
💃 OpenAI’s latest AI Agent that actually does its homework!

OpenAI just dropped a new AI-agent called “deep research”—because, clearly, skimming the surface isn’t cutting it anymore.
This new AI-research tool isn’t here to give you snappy one-liners or quick, overconfident guesses. Nope, deep research is built for people who need serious, well-sourced answers.
Think finance, science, policy, engineering—or even big-ticket purchases like cars and appliances. Basically, anywhere you don’t want to gamble on half-baked info.

Here’s how it works:
Instead of tossing out an instant (and possibly wrong) reply, deep research actually takes its time—anywhere from 5 to 30 minutes—digging through multiple sources to deliver a thorough, well-researched response.
It even lets you upload files and spreadsheets for a deeper dive.
For now, the results are text-only, but OpenAI promises snazzy upgrades—like graphs, images, and sleek data visualizations, to make your reports look extra professional."
Currently, it’s only available to ChatGPT Pro users (limited to 100 queries per month), with Plus, Team, and Enterprise users joining the party soon.
It’s also web-only for now, and if you’re in the U.K., Switzerland, or the European Economic Area, you’ll have to wait a bit longer, as OpenAI is yet to set a release date for those regions.
Now, the million-dollar question: Can you trust it?
Well… kind of.
OpenAI’s powering deep research with a specialized version of its recently announced o3 “reasoning” AI model that's fine-tuned through reinforcement learning (aka, trial and error with digital gold stars for good behavior).
This particular version, according to OpenAI, is optimized for web browsing and data analysis.

It leverages reasoning to sift through mountains of text, images, PDFs, interpret data, cite sources, and even plot graphs.
In testing, the o3 model powering deep research scored 26.6% on something called Humanity’s Last Exam—a brutal academic benchmark designed to trip up even the smartest AI.
Now, that might sound like a failing grade, but when you compare it to rivals like Gemini Thinking (6.2%), Grok-2 (3.8%), and even OpenAI’s own GPT-4o (3.3%), o3’s basically the valedictorian.
But let’s be real—AI is still AI, which means it can and will make mistakes.
Though, according to OpenAI every deep research report will come with clear citations, full documentation, and a summary of how the AI agent arrived at its conclusions, so you can fact-check if needed.
But hey, OpenAI also admits it might struggle with misinformation, mix up credible facts with internet nonsense, misformat reports, and occasionally mess up citations.
And—classic AI move—it’s also not great at admitting when it’s unsure.
So, unless you want to accidentally cite AI-generated nonsense in your next big report, double-checking is still a great idea.
Oh, and about that name… here’s a fun bit of déjà vu:
Just two months ago, Google launched a similar feature with the exact same name.
So, either OpenAI forgot to Google their own product name before launching, or "deep research" is the AI industry’s new favorite buzzword.
Either way, if you’re tired of shallow, half-baked chatbot answers and want something that actually thinks before responding, deep research might be exactly what you’re looking for.
🎁 Referral Radar: From The Automated to The Autonomous Agency 🚀
We’re excited to introduce The Autonomous Agency, the next evolution of The Automated.
After helping 16,000+ subscribers grow with AI-driven insights, we’re now opening up our playbook to help YOU build and scale your personal brand or business with the exact tools and strategies we’ve used.
Here’s how you can get early access and exclusive rewards:
✨ 1 Referral: Unlock exclusive Loom videos where we reveal how we built our processes for creating and managing content, automating workflows, and scaling efficiently.
✨ 3 Referrals: Earn a 50% discount on your first month with The Autonomous Agency. Let’s take your brand or business to the next level with powerful automation!
✨ 5 Referrals: Get a 30-minute one-on-one call where I’ll personally explain how the system works and how you can implement it for your brand or business.
📣 Let’s keep the momentum going—refer your friends and unlock these rewards today! The more you share, the more you gain
🥳 OpenAI rolls out O3-Mini, a new reasoning model with a free ChatGPT version.

OpenAI is proving it’s a company of its word!
Just two weeks after CEO Sam Altman teased the arrival of o3-mini, the new reasoning model is here.
Designed for speed and accuracy, o3-mini is rolling out across ChatGPT and OpenAI’s API services, and for the first time, free ChatGPT users can try a reasoning model—though with rate limits.
Announced during OpenAI’s “12 Days of Ship-mas” in December, o3-mini is built to match o1’s performance in math, coding, and science but responds 24% faster.
It also shows how it arrives at answers instead of just spitting out results—something OpenAI has been refining.
Benchmarks suggest that o3-mini outperforms o1 in coding and reasoning tasks while being cheaper and more efficient.
Developers can access o3-mini via OpenAI’s API services, including Chat Completions, Assistants, and Batch APIs.
Meanwhile, paid users can opt for o3-mini-high, which OpenAI claims is the best coding model in ChatGPT, offering more intelligent responses (albeit with slightly longer wait times).
o3-mini also integrates search, providing answers with links to sources. Neat, huh?
But that’s not all
Free-tier users can now experiment with OpenAI’s reasoning models, just as Microsoft made o1 free for all Copilot users and DeepSeek stirred up competition.
Free ChatGPT users can access o3-mini by selecting the "Reason" feature in the chat bar, with limits similar to GPT-4o’s.
For those on paid plans, OpenAI is tripling daily message limits for Plus and Teams users to 150, while Pro users ($200/month) get unlimited access.
Microsoft is also embracing o3-mini, making it available in Azure OpenAI Service, GitHub Copilot, and GitHub models.
With OpenAI ramping up accessibility and Microsoft pushing AI adoption, the competition in AI reasoning is heating up fast.
Curious to see how o3-mini stacks up, dive into the full post for all the details.
Start learning AI in 2025
Everyone talks about AI, but no one has the time to learn it. So, we found the easiest way to learn AI in as little time as possible: The Rundown AI.
It's a free AI newsletter that keeps you up-to-date on the latest AI news, and teaches you how to apply it in just 5 minutes a day.
Plus, complete the quiz after signing up and they’ll recommend the best AI tools, guides, and courses – tailored to your needs.
🧱Around The AI Block
🦾 How to build AI agents for free.
📈 How to use Claude Artifacts to visualize data.
🔬 OpenAI used this subreddit to test AI persuasion.
❌ Hundreds’ of companies blocks DeepSeek over China data risks.
🧑⚖️ AI startup Perplexity sued for alleged trademark infringement.
🗣️ Apple Intelligence will support more languages starting in April.
🤖 MLCommons and Hugging Face team up to release massive speech dataset for AI research.
🤳 Google’s ‘Ask for Me’ uses AI to call local businesses for you.
🛠️ Trending Tools
Doctranslate.io: Seamlessly translates documents of any type into dozens of languages with precision and ease.
Humanlinker: Helps with personalized prospecting and efficient meeting preparation.
Success AI: An AI-powered platform that provides unlimited email accounts, lead warmups, and turbocharges cold email outreach to help businesses grow.
Designs.ai: An AI-powered platform for instant creation, editing, and scaling of logos, videos, banners, and mockups in 2 minutes.
Julius AI: An AI data analyst to analyze datasets, create visualizations, and even train ML models with only a prompt.
🤖ChatGPT Prompt Of The Day: Validate Data Format.
Validating data format is crucial to maintain data integrity and prevent misinterpretation.
It ensures that data follows expected structures, reducing processing errors and improving efficiency. Clean, well-formatted data leads to more accurate analysis and trustworthy business decisions.
Here's a prompt that can help you with that.
As a Data Analysis Expert, your task is to validate a dataset to ensure it meets a specified [list of conditions]. This involves meticulously checking each data entry against the conditions, such as data type, range values, missing values, and format consistency. You will use statistical and computational methods to identify any discrepancies, outliers, or anomalies in the dataset. Your goal is to ensure the data is clean, accurate, and ready for further analysis or processing. This task requires a keen eye for detail, a solid understanding of data integrity principles, and proficiency in data analysis tools and software. Your findings and any corrections made should be documented in a comprehensive report, outlining the validation process, issues identified, and the steps taken to resolve them, ensuring the dataset's reliability and effectiveness for its intended use.
We've Compiled a List of Over 100 ChatGPT Power Prompts.
This should help streamline your interactions with ChatGPT and get the results you need more efficiently.
Best of all, It's free!

That's all we've got for you today.
Did you like today's content? We'd love to hear from you! Please share your thoughts on our content below👇
What do you think of today's email? |
Your feedback means a lot to us and helps improve the quality of our newsletter.
Reply