
Google Gemini lets you build small apps, turn emails into tasks and work with files by describing what you need. Google Gemini is an AI chat app with a choice of models, thinking levels and custom instructions called skills, as well as a task mode that can work across Google services. Your AI Workflow found that Gemini could create a quote calculator, record incoming leads and draft replies for review. Take a closer look at how to use these options, along with video generation from images and a prompt.
TL;DR Key Takeaways :
- Gemini 4 Hydrogen replaces Flash for quicker tasks, while Argon targets complex work and Carbon is intended for the most complex tasks.
- Hydrogen built a cleaning-business quote calculator in 30 seconds using medium thinking.
- Task mode added two matching email leads to a Google Sheet and saved reply drafts in Gmail.
- Argon processed five videos containing about an hour of footage and produced a milestone document in about five minutes.
- Omni 1.2 generated a 10-second teaser from reference images, with dialogue matching the prompt word for word.
Google Gemini: Models, Apps and Task Tools
Google Gemini combines a chat interface with tools for creating apps, working with files and carrying out tasks across Google services. Its model menu includes replacements and additions for different types of work.
Your model selection determines how Gemini approaches a request, while a separate thinking-level control offers low, medium and high settings. Gemini can also use custom instructions, called skills, when a task calls for them.
Choose a Model for the Work
Gemini 4 Hydrogen replaces Gemini Flash and is intended for quicker tasks. Hydrogen is faster and hallucinates less than its predecessor.
Gemini 4 Argon is aimed at more complex work, including code and analysis, while Gemini 4 Carbon is intended for the most complex tasks. Your AI Workflow recommends Argon for processing large amounts of data and producing an output, but considers it less suited to creative work than Hydrogen.
Build and Share a Productivity App
You can describe a small app in a Gemini prompt, including its inputs and the result it should produce. With Hydrogen set to medium thinking, a request for a cleaning-business quote calculator produced an app in 30 seconds.
The request specified a $40 hourly rate, prices for extras and a 10% returning-client discount. The calculator accepted a five-hour booking, add-ons and the discount, then displayed a total and a client message that could be copied into WhatsApp.
Use Share to copy a link that someone can open to use the calculator. Anyone with the link can also see the app’s contents, so Your AI Workflow warns against including sensitive data.
Check out more relevant guides from our extensive collection on Google Gemini 4 models that you might find useful.
- Gemini 4 Argon AI Super App Launches with Chat and Task Modes
- Hands-on with Gemini 4 Argon and Its 1 Million Tokens
- Gemini 4 Argon vs ChatGPT 6 Astra vs Claude Fable 5.1
- Google Gemini 4 Carbon Beats Claude Opus 5.5 in 16 of 18 Tests
- How Google Gemini 4 Argon Costs 33% Less Than ChatGPT 6.1 to Run
Turn Incoming Emails Into Tasks
Task mode can respond to incoming messages according to instructions in a prompt. For lead collection, you can ask Gemini to watch for “quote” or “inquiry” in an email subject and record the sender’s name, company, request and budget in a Google Sheet.
The automation is meant to check Google Calendar for two open times during the rest of the week and save a Gmail reply offering them, rather than sending it. When Gemini noticed that the leads sheet did not exist, it created an empty one on request. You can view and edit files inside Gemini.
After two matching quote requests arrived, the chat added both clients as rows in the sheet and saved reply drafts in Gmail. One draft offered two available times and could be edited or sent directly.
Process Project Files Into a Shareable Document
Task mode can also carry out a step-by-step request using material you add from Google Drive. You can select Argon and attach a project folder for Gemini to process before it creates an output.
For a milestone breakdown, Argon processed five videos containing about an hour of footage and produced a document in about five minutes. It included a summary, a table of milestones and dates, links to recordings and a detailed feature breakdown. A check against the footage confirmed details that had initially raised concerns about hallucination.
The document opens like a Google Doc, and you can share it in the same way, including by changing access to anyone with the link.
Create Skills for Repeat Tasks
Skills are custom instructions for particular tasks that you can create directly in a Gemini chat. To request an email-responder skill, you can ask Gemini to analyse the previous two weeks of emails and identify what to say, when to say it and the writer’s style.
A resulting skill specifies when to use it, and Gemini can automatically load a relevant skill without a separate reminder. In a task for sponsored-video or partnership messages, Gemini loaded the email-responder skill and drafted a reply in the requested style with the correct details for review.
That automation needs keywords to identify the messages; it will not work with every email. You can then review and send the draft rather than having it dispatched automatically.
Generate Video From Images and a Prompt
Omni 1.2 is a video model that accepts reference images alongside a text prompt. It is intended to follow reference images and instructions closely and recreate requested dialogue.
A request using separate hoodie and logo images produced a 10-second launch teaser whose dialogue followed the prompt word for word. One detail did not match the scene: the person asking about the hoodie never looked at it, making the question less clear. Your AI Workflow suggests that might be fixable through chat.
Start by choosing a model for the work: Hydrogen is intended for quicker requests, while Argon is aimed at more complex tasks. You can share an app with a link, but leave out sensitive information because anyone with the link can see its contents. For anyone using Google Gemini to build a small app, begin with a prompt describing its inputs and the result you want.
Media Credit: Your AI Workflow
Disclosure: Some of our articles include affiliate links. If you buy something through one of these links, Geeky Gadgets may earn an affiliate commission. Learn about our Disclosure Policy.