This is Hamamoto from TIMEWELL.
In January 2025, OpenAI announced "Operator," ushering in an era where AI could autonomously operate a browser.
Just seven months later, in August 2025, Operator was discontinued and its capabilities were integrated into "ChatGPT Agent"—a next-generation agent that fuses browser operation, deep research, and conversational AI into one.
This article covers how ChatGPT Agent works, its practical use cases, and what the autonomous AI era of 2026 will look like.
ChatGPT Agent: 2026 Latest Information
| Item | Details |
|---|---|
| Operator Announced | January 23, 2025 |
| Operator Discontinued | August 31, 2025 |
| ChatGPT Agent Integration | Second half of 2025 |
| Underlying Model | CUA (Computer-Using Agent) |
| Available Plan | ChatGPT Pro ($200/month) |
What Was Operator?
The January 2025 Shock
Operator was an AI agent announced by OpenAI on January 23, 2025. It could autonomously operate a web browser to perform tasks such as:
Operator's Capabilities:
- Automating form entry
- Executing online orders
- Scheduling reservations
- Browsing websites and gathering information
- Automating repetitive tasks
The CUA (Computer-Using Agent) Model
The core technology behind Operator was a new model called CUA (Computer-Using Agent).
How CUA Works:
- Screenshot Capture: Adds a screenshot of the computer screen to the model's context
- Reasoning: Uses Chain-of-Thought to determine the next step
- Action Execution: Performs clicks, scrolling, and typing
- Iteration: Continues until the task is complete or user input is required
This loop enabled AI to "see," "think," and "operate" websites in a human-like way.
Why Was Operator Discontinued?
Operator was discontinued on August 31, 2025, and integrated into ChatGPT Agent.
Reasons for Integration:
- Eliminate overlapping functionality
- Unify the user experience
- Provide more advanced agent capabilities
- Merge with Deep Research
Looking for AI training and consulting?
Learn about WARP training programs and consulting services in our materials.
ChatGPT Agent — The Fusion of Three Capabilities
A Unified Agent
ChatGPT Agent fuses three capabilities:
| Capability | Previous Form | Integration in ChatGPT Agent |
|---|---|---|
| Browser Operation | Operator | Runs on a virtual computer |
| Information Synthesis | Deep Research | Seamlessly integrated |
| Conversational AI | ChatGPT | Natural dialogue for instructions |
Running on a Virtual Computer
ChatGPT Agent executes tasks on its own "virtual computer."
Features of the Virtual Computer:
- An environment independent of the user's PC
- A secure execution space
- Data sharing across multiple tools
- Fluid switching between reasoning and action
This enables seamless processing of complex workflows from start to finish.
Practical Use Cases
1. Automating Business Research
Traditional Process:
- Manually browse multiple websites
- Collect information by hand
- Compile into a spreadsheet
- Write a report
Process with ChatGPT Agent:
- Instruct: "Research the latest products from five competitors and compile them into a spreadsheet"
- The agent automatically executes the entire process
- Receive the completed report
2. Automating Reservations and Orders
Use Cases:
- Restaurant reservations
- Hotel reservations
- Online shopping
- Ticket purchases
- Government service applications
3. Automating Routine Tasks
Use Cases:
- Entering expense reimbursement data
- Registering time and attendance in HR systems
- Entering customer information into a CRM
- Creating and sending periodic reports
Security and Safety Measures
Multi-Layered Protection
Because ChatGPT Agent executes tasks autonomously, security is especially important.
Safety Measures:
- User Authentication: Requires confirmation before critical actions
- Access Control: Restricts access to external APIs
- Data Encryption: Encrypts data in transit and at rest
- Stop Function: Can halt tasks at any time
- Audit Logs: Records all actions
Use in Finance and Healthcare
Even in fields with high security requirements, adoption is advancing with appropriate safeguards.
Finance:
- Explicit permission required for access to account information
- Double confirmation before executing transactions
Healthcare:
- Restricted access to patient information
- HIPAA-compliant operations
Then and Now: The Evolution of OpenAI Agents
| Item | Then (January 2025, Operator Launch) | Now (January 2026) |
|---|---|---|
| Agent | Operator (standalone) | ChatGPT Agent (unified) |
| Browser Operation | Operator-exclusive | Integrated execution on virtual computer |
| Deep Research | Separate feature | Integrated into Agent |
| Availability | Pro only (US only) | Pro/Team/Enterprise |
| CUA Model | Initial version | Improved version |
| Operator Status | Active | Discontinued (August 2025) |
2026 Outlook
Evolution Beyond the Browser
In 2026, ChatGPT Agent is expected to evolve beyond web browser capabilities.
Future Development Directions:
- Operating desktop applications
- Integration with mobile apps
- Controlling IoT devices
- Integration with the physical world (robotics)
Expansion to More Affordable Plans
The agent features currently limited to Pro ($200/month) may expand to more affordable plans.
Anticipated Expansion:
- Feature-limited availability on Plus ($20/month)
- Full deployment for Team/Enterprise
- Expanded API access
Proactive Agents
In the future, "proactive agents" that autonomously initiate tasks without waiting for user instructions may emerge.
Examples of Proactive Capabilities:
- Automatic execution of periodic data analysis
- Automatic scheduling suggestions
- Automated monitoring of market changes with alerts
- Automatic email draft generation
Enterprise Deployment Strategy
Phased Adoption Approach
Phase 1: Pilot
- Test with limited use cases
- Measure ROI
- Assess risk
Phase 2: Departmental Rollout
- Expand to areas with proven results
- Employee training
- Build governance frameworks
Phase 3: Company-Wide Deployment
- Full integration into business processes
- Expand automation scope
- Continuous improvement
Key Considerations for Deployment
- Security Policy: Control access to confidential information
- Compliance: Address industry regulations
- Employee Communication: Share the value of working alongside AI
- Monitoring: Oversee agent behavior
Summary
ChatGPT Agent has been reborn as a more integrated and powerful AI agent following the discontinuation of Operator.
Key Takeaways:
- Operator was discontinued in August 2025 and integrated into ChatGPT Agent
- The CUA (Computer-Using Agent) model enables browser operation
- Fusion of browser operation + deep research + conversational AI
- Tasks run safely on a virtual computer
- 2026 expected to bring evolution beyond the browser and expansion to more affordable plans
AI agents are evolving from a "convenient tool" into a "transformative force for business." 2026 will be the year when how strategically companies deploy this technology determines their competitive edge.
Early adopters can accumulate not just operational efficiency, but an intangible asset in the form of know-how for working alongside AI. Now is the time to seriously consider deploying ChatGPT Agent.
Related Articles
- From Full-Time to Part-Time: Life After Two Maternity Leaves and How My View of Work Changed | TIMEWELL
- Before Paternity Leave (Part 2): Three Things You Absolutely Must Do to Take Leave During Busy Season
- Staying True to the Field: How the 5th-Generation Head of a Construction Firm Found His Own Way | Fujita Construction
