Collate WHO data into a multi-tab spreadsheet
2025 has seen agents coming in thick and fast, but when ChatGPT launches Agent mode, it's time to take notice. Task: 2025 has seen agents coming in thick and fast, but when ChatGPT launches Agent mode, it's time to take notice. We've officially hit the tipping point. Every major AI company has launched autonomous agents in 2025 - OpenAI, Anthropic, Microsoft, Google, Amazon, and Meta are all racing to give AI the ability to act, not just answer. After two years of impressive conversations, AI is finally stepping out of the chat box to book your flights, manage your projects, and handle your shopping. We're witnessing AI's "iPhone moment", the shift from experimental novelty to essential productivity tool. Agents are not only going to change how everyone interacts with computers. They’re also going to upend the software industry, bringing about the biggest revolution in computing since we went from typing commands to tapping on icons. Bill Gates - The Future Of Agents (2023) OpenAI's July launch of ChatGPT Agent Mode is the breakthrough that matters most. Why? Availability. While competitors like Manus AI and Genspark target power users, ChatGPT Agent brings autonomous AI to 100+ million existing users for just $20/month. No waitlists, no technical setup, just click "Agent mode" and start delegating. This democratisation changes everything. Agents aren't just for tech teams or power users anymore; they're for every marketing manager, sales director, and productivity worker. Ask ChatGPT to “Craft a 5-day foodie & winery itinerary for Margaret River using TripAdvisor” and it browses deep into Trip Advisor to research wineries, restaurants, and destinations, creating day-by-day schedules (and could actually make bookings with a bit more prompting). What used to take hours (or even days) now happens in minutes. Beyond product recommendations, ChatGPT Agent aggregates data from retailers, and expert reviews to find exactly what you need. In this case, ask ChatGPT agent mode to “Find and and buy the top-rated noise-cancelling true-wireless earbuds on JB Hi-Fi” and it automatically goes away, starts researching and adding to basket, dramatically cutting down research time, and disrupting the buying journey. For presentations, you can ask ChatGPT Agent mode to “Develop a go-to-market presentation for a luxury hospitality brand entering a new international market.” In minutes, it produces a polished slide deck covering target audience insights, brand positioning, competitive landscape, distribution strategy, and marketing channels. This used to require weeks of time and now it’s done in minutes. Collating data has just become much easier too. Ask ChatGPT Agent mode to “Retrieve the most recent publicly available data from the World Health Organisation (WHO)” and watch it auto-navigate the site, extracts relevant data, and organises it into a clean, multi-tabbed spreadsheet sorted by indicators, countries, and trends. ChatGPT Agent Mode can also connect to enterprise tools like Google Drive, SharePoint, Notion, GitHub, and Teams via ChatGPT Connectors. This means agents you run can autonomously pull relevant files, scan calendars, summarise documents, and cross-reference data from Gmail, Linear, or HubSpot, all in one seamless workflow. Limited website access: ChatGPT Agents frequently fail to interact with many websites due to login issues, unsupported browser states, or incompatibility with dynamic content and single-page applications. Limited website access: ChatGPT Agents frequently fail to interact with many websites due to login issues, unsupported browser states, or incompatibility with dynamic content and single-page applications. Security and privacy vulnerabilities: Agents can be susceptible to prompt injection and other hacks, and lack clear audit trails, making it risky for sensitive or critical actions. Don’t let them do your banking yet! Security and privacy vulnerabilities: Agents can be susceptible to prompt injection and other hacks, and lack clear audit trails, making it risky for sensitive or critical actions. Don’t let them do your banking yet! Inconsistent usability: Outputs from Agents often need reformatting or correction, and some tasks are easier or faster to do manually (e.g. booking an Uber, one of the OpenAI examples, can take over 15 minutes to execute!) Inconsistent usability: Outputs from Agents often need reformatting or correction, and some tasks are easier or faster to do manually (e.g. booking an Uber, one of the OpenAI examples, can take over 15 minutes to execute!); tools: ChatGPT Agent. Human role: The original post does not say how a person stayed involved.; limitations: Agents face limited website access due to login issues or dynamic content incompatibility. Security vulnerabilities include susceptibility to prompt injection and a lack of clear audit trails, making sensitive actions risky. Inconsistent usability means manual tasks may sometimes be faster than agent execution..
Tools
Published on Agent Work Atlas:
Overview
2025 has seen agents coming in thick and fast, but when ChatGPT launches Agent mode, it's time to take notice.
Workflow
2025 has seen agents coming in thick and fast, but when ChatGPT launches Agent mode, it's time to take notice. We've officially hit the tipping point. Every major AI company has launched autonomous agents in 2025 - OpenAI, Anthropic, Microsoft, Google, Amazon, and Meta are all racing to give AI the ability to act, not just answer. After two years of impressive conversations, AI is finally stepping out of the chat box to book your flights, manage your projects, and handle your shopping. We're witnessing AI's "iPhone moment", the shift from experimental novelty to essential productivity tool. Agents are not only going to change how everyone interacts with computers. They’re also going to upend the software industry, bringing about the biggest revolution in computing since we went from typing commands to tapping on icons. Bill Gates - The Future Of Agents (2023) OpenAI's July launch of ChatGPT Agent Mode is the breakthrough that matters most. Why? Availability. While competitors like Manus AI and Genspark target power users, ChatGPT Agent brings autonomous AI to 100+ million existing users for just $20/month. No waitlists, no technical setup, just click "Agent mode" and start delegating. This democratisation changes everything. Agents aren't just for tech teams or power users anymore; they're for every marketing manager, sales director, and productivity worker. Ask ChatGPT to “Craft a 5-day foodie & winery itinerary for Margaret River using TripAdvisor” and it browses deep into Trip Advisor to research wineries, restaurants, and destinations, creating day-by-day schedules (and could actually make bookings with a bit more prompting). What used to take hours (or even days) now happens in minutes. Beyond product recommendations, ChatGPT Agent aggregates data from retailers, and expert reviews to find exactly what you need. In this case, ask ChatGPT agent mode to “Find and and buy the top-rated noise-cancelling true-wireless earbuds on JB Hi-Fi” and it automatically goes away, starts researching and adding to basket, dramatically cutting down research time, and disrupting the buying journey. For presentations, you can ask ChatGPT Agent mode to “Develop a go-to-market presentation for a luxury hospitality brand entering a new international market.” In minutes, it produces a polished slide deck covering target audience insights, brand positioning, competitive landscape, distribution strategy, and marketing channels. This used to require weeks of time and now it’s done in minutes. Collating data has just become much easier too. Ask ChatGPT Agent mode to “Retrieve the most recent publicly available data from the World Health Organisation (WHO)” and watch it auto-navigate the site, extracts relevant data, and organises it into a clean, multi-tabbed spreadsheet sorted by indicators, countries, and trends. ChatGPT Agent Mode can also connect to enterprise tools like Google Drive, SharePoint, Notion, GitHub, and Teams via ChatGPT Connectors. This means agents you run can autonomously pull relevant files, scan calendars, summarise documents, and cross-reference data from Gmail, Linear, or HubSpot, all in one seamless workflow. Limited website access: ChatGPT Agents frequently fail to interact with many websites due to login issues, unsupported browser states, or incompatibility with dynamic content and single-page applications. Limited website access: ChatGPT Agents frequently fail to interact with many websites due to login issues, unsupported browser states, or incompatibility with dynamic content and single-page applications. Security and privacy vulnerabilities: Agents can be susceptible to prompt injection and other hacks, and lack clear audit trails, making it risky for sensitive or critical actions. Don’t let them do your banking yet! Security and privacy vulnerabilities: Agents can be susceptible to prompt injection and other hacks, and lack clear audit trails, making it risky for sensitive or critical actions. Don’t let them do your banking yet! Inconsistent usability: Outputs from Agents often need reformatting or correction, and some tasks are easier or faster to do manually (e.g. booking an Uber, one of the OpenAI examples, can take over 15 minutes to execute!) Inconsistent usability: Outputs from Agents often need reformatting or correction, and some tasks are easier or faster to do manually (e.g. booking an Uber, one of the OpenAI examples, can take over 15 minutes to execute!)
Reported result
The source describes the intended workflow but does not provide specific metrics on execution time or accuracy for this task. It notes that outputs often need reformatting or correction. A measurable acceptance check would be to compare the agent's extracted data against the official WHO source for a specific indicator to determine error rates.
Human role
The original post does not say how a person stayed involved.
Limitations
Agents face limited website access due to login issues or dynamic content incompatibility. Security vulnerabilities include susceptibility to prompt injection and a lack of clear audit trails, making sensitive actions risky. Inconsistent usability means manual tasks may sometimes be faster than agent execution.
Try it
I want to test ChatGPT Agent Mode for data collation. Please execute the following steps: 1. Navigate to the World Health Organisation (WHO) website. 2. Retrieve the most recent publicly available data for [replaceable input: specific health indicator, e.g., Malaria cases]. 3. Organize this data into a multi-tabbed spreadsheet sorted by [replaceable input: countries] and [replaceable input: years/trends]. 4. Do not save or publish the file yet. 5. Present the raw output and the proposed spreadsheet structure to me for review. I will verify the data accuracy against the source before confirming completion. If the agent fails to navigate due to dynamic content, report the error immediately.
Source reports are not independently verified. Publication date is the catalog publication date, not the original article date.