Use Gemini 3.5 Flash to Automate On-Screen Tasks
Delegate repetitive digital clicks and navigation to Gemini's computer use model.
Utilize Gemini 3.5 Flash's 'computer use' capabilities to automate multi-step digital workflows that require interacting with standard software interfaces.
The Scenario
You need to perform a repetitive digital task, such as scraping specific data from a web portal and entering it into a internal database or sheet. Instead of clicking through yourself, delegate the navigation to the AI.
Before & after
You would manually click through multiple tabs, copy data points into a spreadsheet, and format the final document yourself. Time cost: 45–60 minutes.
Run a 'Computer Use' agent to navigate local files, web browsers, and apps to gather data and generate a report autonomously. Time cost: 2–4 minutes of oversight.
The Prompt
Using Gemini 3.5 Flash with computer use enabled: 'Open the [WEBSITE_URL], find the pricing for the three main tiers, enter them into a new Google Sheet, and calculate the annual cost for each.'
Gemini 3.5 Flash's computer use capability allows the model to interact with interfaces much like a human would, performing multi-step digital tasks.
Source
Official Gemini news and updates | Google Blog"Introducing computer use in Gemini 3.5 Flash"