Published Date:
February 4, 2025

AI Agents Are Now at Your Fingertips

Explore how OpenAI’s new "agentic" capabilities, Operator and Deep Research, perform real business tasks and if they are worth the $200 price tag.
By
Daan van Rossum
Founder & CEO, FlexOS

I Paid $200/Month for ChatGPT to Discover the Future of AI.

Last week, I did something that might seem crazy: I upgraded to ChatGPT's $200/month plan—10x the regular $20 subscription.

Why?

Because I wanted to see if OpenAI's latest "agentic" features were worth the hype, and more importantly, what they mean for the business leaders in our community.

After testing some of the PRO features, I'm convinced that what I've discovered is worth taking a closer look at together.

About ChatGPT PRO – And Why It May Be Worth $200/Month

​ChatGPT PRO​ is OpenAI’s premium subscription tier, offering unlimited use of its newer 'reasoning’ models and enhanced capabilities, including video generator ​Sora​.

Of course, its introduction ​quickly sparked debate​ about whether this would ever be worth the $200 price tag.

Well, I guess there was only one way to find out:

Because while the standard version of ChatGPT is already powerful, the PRO plan introduces "agentic" capabilities—allowing AI to take real actions, make decisions, and refine its work in a way that goes beyond simple input-output responses.

For most of us, there's little need for any of this.

But it does give us a good glimpse into the future, especially what I discovered about two flagship features: Operator and Deep Research.

Operator: AI That Takes Real Action

We've all gotten comfortable with AI as a tool—input goes in, output comes out. It's been an incredible assistant, but a fairly passive one.

Enter Operator.

​ChatGPT Operator​ builds on something we've seen in Claude Computer Use (see our guide ​here​) before: it can navigate websites and interfaces all by itself.

It's like magic.

Instead of just responding, AI is beginning to act on its own, handling workflows we had to get involved in ourselves until now.

(Screenshot: I asked to create a new Google Sheet with the top 10 global companies by revenue. It finalized the research quickly, navigated to Google Sheets, and asked me to log in.)

In demos like the one below, OpenAI shows how you could use Operator to book concert tickets or do online shopping for you, but the business implications are much more interesting.

Think about all the cloud-based systems that don't offer no or poor AI assistance. Using Operator, ChatGPT can now log in, navigate, and execute tasks on these sites.

It could upload invoices to Xero for accounting, log in to your CRM to update customer records, or pull reports from your HR system—without you lifting a finger.

Deep Research: Your Personal Analyst

​Deep Research​ tackles complex topics with in-depth, multi-step analysis.

It gathers data from various online sources to generate truly comprehensive reports—I just got a 26-page analysis on membership best practices.

This process, powered by OpenAI's ​new o3 "reasoning model,"​ takes the AI between 5 to 30 minutes, and aims to match the output of a human research analyst.

And this one genuinely surprised me.

Unlike traditional AI, which immediately starts working from a prompt, Deep Research begins by asking thoughtful, clarifying questions.

When I asked it to compile and analyze a set of case studies about B2B creators, it asked whether I wanted to focus on specific industries, formats, or time frames:

It then went off to do the research - for over 6 minutes - keeping me updated at each step. This was truly remarkable to see - similar to seeing ChatGPT itself for the first time.

It felt less like giving commands to AI and more like working with a senior analyst.

The Major Shift: From Assistant to Agent

The evolution of AI from assistant to independent operator is becoming clearer, and I find it helpful to think about this progression in four key layers:

  • Regular AI: A linear input/output model that responds to each prompt.
  • AI Assistants: Specialized tools that remember your preferences, like GPTs.
  • AI Automation: Systems that trigger responses automatically from specific events.
  • True AI Agents: Plan, decide, and coordinate tasks independently, exactly what Operator and Deep Research deliver.

And while the $200 price tag isn't justified for most users right now, what I've seen gives us a clear preview of where AI is heading.

As always, I'd love to hear your thoughts—would you ever drop the $200 for AI agents?

Lead with AI - Icon

SPONSORED BY GUIDDE

Create How-To Video Guides in Minutes With AI

Want to create helpful FAQs for customers or SOPs for teams? Guidde can help you create shareable guides in less time.

​Guidde is an AI that records you doing a task and automatically creates a professional video with instructions:

  • Screen-record your task (or upload a PDF)
  • Choose from 100+ voices and 35 languages for captions
  • Upload logos, colors, fonts, & more for branding

In just a few minutes, you’ll have a video guide ready to share everywhere. You can even track performance to optimize engagement.

The best part? You can try Guidde for free.

‍

>> EXPLORE GUIDDE NOW

Lead with AI - Icon

Every week, I highlight a real-world AI use case to spark ideas and inspirations for how you can implement AI in your team and business. For this week:

Lyft Resolves Driver Requests 87% Faster with Claude AI

Image of Lyft chatbot

AI implementation doesn’t have to be complex to deliver results. Take Lyft.

By integrating Anthropic’s Claude AI to handle common driver inquiries, the company reduced resolution times by 87%, freeing human specialists to focus only on complex cases.

Beyond this implementation, Lyft uses generative AI to enhance engineering productivity, with 25% of its code now AI-generated.

The company is also exploring broader AI integrations to streamline processes, boost productivity, and improve both driver and rider experiences.

​>> Read more information here. ​

Lead with AI - Icon

News & Updates

  • ***LIVE Masterclass this Thursday, Feb 13: Explore and apply ChatGPT’s newest features in practical ways. ​RSVP here​ (with my name in the referral field.)

‍

  • AI is being adopted faster than the internet or PCs, but its productivity gains lag. DeepSeek even shakes up the market recently. This Bloomberg article explores if AI will fuel growth, falter, or disrupt—especially with DeepSeek shaking up the market.
  • In his latest blog, Sam Altman share his thoughts on how AI’s falling costs, virtual co-worker agents, and future access to humanity’s collective intelligence could transform the economy and unlock new opportunities. Read it here.
  • Google launched its ultra-affordable Gemini Flash 2.0 with a massive 1-million-token context window and Gemini 2.0 Pro (Experimental) with an industry-leading 2-million-token window for coding and advanced reasoning.
  • Google has updated its AI principles, quietly removing its 2018 pledge not to apply AI to weapons or surveillance, citing the need to support national security efforts in a competitive global AI landscape.
  • SoftBank is investing $3 billion annually to OpenAI to develop and implement “Cristal Intelligence” across its companies, with plans to automate 100 million workflows and launch a joint venture to bring customized AI solutions to Japan’s largest enterprises.
  • GitHub Copilot now features Agent Mode for automated debugging and multi-step task execution. Its Project Padawan also previews AI agents that can autonomously generate and test pull requests, bringing hands-free coding closer to reality.‍
  • Elon Musk's DOGE is building an AI chatbot called GSAi to support 12k employees in the US government by automating workflows and analyzing procurement data, part of Trump’s AI-first initiative to modernize federal operations and cut costs.
Lead with AI - Icon

From The Lead with AI Community

Every day, ​Lead with AI PRO members​ discuss practical ways to benefit from AI in their work and organizations. This week's highlights include:

Don't want to miss more insights and conversations like these? Then it's time to upgrade to PRO:

Join The Leading Business AI Community
Lead with AI - Icon

If you made it this far, reply and tell me what you'd love AI to take over in your daily workflow.

Also, please forward this newsletter to a colleague and ask them to subscribe.

If you have any other questions or feedback, just reply here or inbox me.

See you next week,

Daan van Rossum - Lead with AI

Daan van Rossum​

Host, Lead with AI

Daan van Rossum

Founder & CEO, FlexOS

I Paid $200/Month for ChatGPT to Discover the Future of AI.

Last week, I did something that might seem crazy: I upgraded to ChatGPT's $200/month plan—10x the regular $20 subscription.

Why?

Because I wanted to see if OpenAI's latest "agentic" features were worth the hype, and more importantly, what they mean for the business leaders in our community.

After testing some of the PRO features, I'm convinced that what I've discovered is worth taking a closer look at together.

About ChatGPT PRO – And Why It May Be Worth $200/Month

​ChatGPT PRO​ is OpenAI’s premium subscription tier, offering unlimited use of its newer 'reasoning’ models and enhanced capabilities, including video generator ​Sora​.

Of course, its introduction ​quickly sparked debate​ about whether this would ever be worth the $200 price tag.

Well, I guess there was only one way to find out:

Because while the standard version of ChatGPT is already powerful, the PRO plan introduces "agentic" capabilities—allowing AI to take real actions, make decisions, and refine its work in a way that goes beyond simple input-output responses.

For most of us, there's little need for any of this.

But it does give us a good glimpse into the future, especially what I discovered about two flagship features: Operator and Deep Research.

Operator: AI That Takes Real Action

We've all gotten comfortable with AI as a tool—input goes in, output comes out. It's been an incredible assistant, but a fairly passive one.

Enter Operator.

​ChatGPT Operator​ builds on something we've seen in Claude Computer Use (see our guide ​here​) before: it can navigate websites and interfaces all by itself.

It's like magic.

Instead of just responding, AI is beginning to act on its own, handling workflows we had to get involved in ourselves until now.

(Screenshot: I asked to create a new Google Sheet with the top 10 global companies by revenue. It finalized the research quickly, navigated to Google Sheets, and asked me to log in.)

In demos like the one below, OpenAI shows how you could use Operator to book concert tickets or do online shopping for you, but the business implications are much more interesting.

Think about all the cloud-based systems that don't offer no or poor AI assistance. Using Operator, ChatGPT can now log in, navigate, and execute tasks on these sites.

It could upload invoices to Xero for accounting, log in to your CRM to update customer records, or pull reports from your HR system—without you lifting a finger.