tech · 2026-07-10
OpenAI Launches ChatGPT Work Agent
2e74a4af-eadd-4ab9-8f66-ed950ecd57b4
OpenAI unveiled ChatGPT Work, a GPT-5.6 powered agent that runs for hours across apps, files, and workflows to complete complex projects autonomously.This marks OpenAI's biggest push into workplace automation as it prepares for an IPO, directly competing with Microsoft Copilot and Google's Gemini agents.Knowledge workers, IT teams, and SaaS companies face disruption as the tool generates documents, spreadsheets, presentations, and web apps from a single prompt.
What exactly can ChatGPT Work do on its own?
ChatGPT Work operates across multiple apps and files simultaneously. A user can give a broad goal like 'prepare a quarterly review,' and the agent gathers data from connected tools, builds a spreadsheet, drafts a presentation, and assembles a document. Unlike standard ChatGPT, it runs continuously for hours without needing repeated prompts.
How does it differ from standard ChatGPT?
Standard ChatGPT responds to one prompt at a time and loses context between sessions. ChatGPT Work maintains a persistent workspace, chaining multiple sub-tasks autonomously. Think of it less like a chatbot and more like a junior employee. Google's Project Mariner attempted similar multi-step browsing but lacked file-creation depth.
Can it access third-party apps like Slack?
OpenAI confirmed it connects to apps, files, and workflows, though specific integrations weren't named. Competitors like Microsoft Copilot plug into Office 365, Teams, and Outlook natively. For ChatGPT Work to compete, it likely needs connectors for tools like Slack, Google Workspace, and Salesforce, where enterprise workflows actually live.
What does GPT-5.6 add over GPT-4o?
GPT-5.6 improves multi-step reasoning and sustained task execution. Earlier models like GPT-4o could plan tasks but struggled with long execution chains. The upgrade lets the agent maintain coherence across hours-long projects. For comparison, GPT-4o's context window held ~128K tokens. GPT-5.6 reportedly handles far longer operational sequences.
Why launch this just before OpenAI's IPO?
Workplace automation tools are a key revenue growth story for IPO investors. OpenAI competes directly with Microsoft's Copilot, which already generates ~$10Bn in annualized revenue. Launching ChatGPT Work now positions OpenAI to demonstrate enterprise traction to potential shareholders, justifying its reported $300Bn valuation.
Does enterprise AI justify a $300Bn valuation?
The valuation depends on capturing recurring enterprise subscriptions. Salesforce earns ~$35Bn/year selling workflow tools to businesses. If OpenAI converts even 5% of global enterprise software spend toward AI agents, the math works. But profitability remains unproven. OpenAI reportedly lost ~$5Bn in 2024 despite $3.7Bn in revenue.
How does Copilot's head start affect OpenAI?
Microsoft Copilot is embedded in Office 365, used by over 400Mn people. That distribution advantage is enormous. OpenAI must convince enterprises to adopt a separate tool rather than use what's already built into their existing Microsoft stack. Copilot's integration with Teams and SharePoint creates switching costs ChatGPT Work must overcome.
Could this tool accelerate OpenAI's revenue?
OpenAI's current revenue comes mostly from consumer subscriptions and API usage. Enterprise tools like ChatGPT Work typically command $25 to $50 per user per month. With ~200Mn weekly ChatGPT users, converting even a fraction to paid Work subscriptions could add billions. Anthropic's Claude enterprise tier follows a similar per-seat pricing model.
Which workers face the most disruption?
Roles heavy on document assembly, data consolidation, and presentation creation are most exposed. McKinsey estimates ~30% of tasks in office-based jobs could be automated by AI agents. Analysts, junior consultants, executive assistants, and operations coordinators at firms like Deloitte or TCS would see significant workflow changes.
Which Indian IT firms are most exposed?
Indian IT services firms like TCS, Infosys, and Wipro derive significant revenue from document processing, report generation, and data consolidation projects. TCS alone employs ~600K workers. If clients automate these tasks with AI agents, demand for lower-tier service contracts could shrink, pressuring margins on routine engagements.
Could this replace offshore back-office roles?
Back-office roles in areas like financial reconciliation, invoice processing, and report compilation are prime targets. India's BPO sector employs ~1.5Mn workers in such functions. Companies like Genpact and WNS already use AI tools internally. ChatGPT Work could let clients bypass outsourcing entirely for simpler document-assembly workflows.
How might workers adapt rather than lose jobs?
Workers who learn to supervise and audit AI agent output gain value. The role shifts from creating spreadsheets to verifying them. Coursera reports a 3x increase in enrollment for AI-related courses since 2024. Firms like Accenture now require all new hires to complete AI fluency training, reframing the skillset rather than eliminating the role.
Source: thehindubusinessline.com