Get the newsletter
Google · AI / ML · Filed Feb 12, 2025 · USPTO published Aug 13, 2026 · Aug 13 release

Google Patent Turns a Single Prompt Into Documents With Text, Images, and Charts

Right now, building a document with text, charts, and images means bouncing between tools. Google is patenting a way to do all of that from a single typed request.

Network system connecting client devices and servers for generating documents. Drawing from patent filing US 2026/0236891 A1.
Network system connecting client devices and servers for generating documents.All 4 drawings ↓
Publication number
US 2026/0236891 A1(US20260236891A1)
Applicant
Google LLC
Filed · Published
Feb 12, 2025 · Aug 13, 2026
Inventors
Gregory George Galante, Andrew James Motika, Behnoosh Hariri, William Joshua Billingham, Heyang Wei, Matthew Robert Hughes, Merwan Vishnu Hade, Nicholas Joseph Pesce, Sunny Yelu Tian, Weihao Lin, Fenil B. Shah
Status when we published
Company responded; back with the examiner
Document
7 claims
US classification
705/7.18
Examiner
WARNER, PHILIP N (Art Unit 3624)

How Google's single-prompt doc builder actually works

Today, if you want a business report with written analysis, a chart, and some supporting images, you assemble each piece yourself, switching between apps or AI tools for each job. Google wants to change that with a system where you type one request and a coordinating AI handles every piece automatically.

Here's how it plays out for you: you type something like "make a quarterly summary with a sales chart and a header image" into a Google Workspace app. A master AI reads that, figures out which specialist AI tools to call (one for text, one for charts, one for images), collects all the results, and stitches them together into a finished document, formatted and ready to use.

The whole thing arrives in one shot, already formatted for the app you're working in, rather than as a pile of separate files you'd have to arrange yourself.

From the filing · CLAIM 1
determine that the content is to include a first content item of a first content type and a second content item of a second content type; identify a second AI model trained to generate content items of the first content type and a third AI mode trained to generate content items of the second content type …

Translation: The system decides what kind of media you need and picks the right specialized AI tools to create each part.

Inside Google's three-model content pipeline

The patent describes a three-layer system built around a coordinating AI model (the "first AI model" in Google's language) that acts as a project manager.

When you submit a text prompt, that coordinator does four things:

  • Decides what kinds of content the finished output needs (text, images, data visualizations, or others)
  • Routes your prompt to the appropriate specialist AI models, one per content type
  • Collects each specialist's output in a shared abstract format (a neutral internal representation that isn't tied to any specific app yet)
  • Combines all the pieces into one unified result, then converts that into the file format your productivity app actually uses

The abstract format is a key design choice. By keeping everything in a common internal structure during assembly, the system can mix content types freely before committing to any specific output format, like a Google Doc or Slides deck. That conversion step happens only at the end.

The patent covers this as a server-side pipeline, meaning the heavy AI work happens in Google's cloud and the finished, app-compatible document is pushed to your device through the app's existing interface.

From the filing · THE ABSTRACT
Methods for generating high fidelity multi-modal content using a unified AI architecture are provided. A textual prompt including a request to generate content is received via a user interface (UI) of a productivity application.

Translation: Google is building a way for office software to create complex documents containing text and images from a single command.

What this means for Google Workspace users

For everyday Google Workspace users, this could mean the difference between spending an hour assembling a presentation and getting a draft in seconds. The patent targets productivity apps specifically, so Docs, Slides, and similar tools are the obvious candidates for this kind of one-prompt generation.

The broader picture here is that Google is trying to own the coordination layer, the AI that decides which other AIs to call, not just the individual generation tools. That architectural bet shows up clearly in Big Tech patent news covering the AI-productivity space, where the race is increasingly about who controls the orchestration, not just who has the best text or image model.

Source: US 2026/0236891 A1 at the USPTO (PDF). Full patent text and figures are in the official publication.

Get our take in your Top Stories

Liked this breakdown? Add Patentlyze as a preferred source on Google, and our plain-English take shows up more often in your Top Stories the next time Google patent news breaks.

There are more where this came from

We read every patent application Big Tech publishes and send you the ones worth knowing. Plain English, free, every week.

The drawings · 4 sheets

4 drawing sheets from US 2026/0236891 A1 · click any drawing to enlarge

System architecture of document generation, drawing sheet 1 from US 2026/0236891 A1
Sheet 1 System architecture of document generation
Flowchart of content generation steps, drawing sheet 2 from US 2026/0236891 A1
Sheet 2 Flowchart of content generation steps
Steps for combining multiple content types, drawing sheet 3 from US 2026/0236891 A1
Sheet 3 Steps for combining multiple content types
Computer hardware system architecture, drawing sheet 4 from US 2026/0236891 A1
Sheet 4 Computer hardware system architecture
Patent filing page

Watchlists this filing belongs to

All watchlists →