# AI presentation: what Gemini in Google Slides means for speakers

Author: Marin Niehues, Stage Starter · Published: 2026-07-29
Topic: Talks & Stage (https://www.stage-starter.de/en/blog/topic/talks-and-stage/)
Source: https://www.stage-starter.de/en/blog/ai-presentation-google-slides/

You can create an AI presentation directly in Google Slides since 23 July 2026, because on that day Google announced on the [Google Workspace Blog](https://workspace.google.com/blog/product-announcements/create-on-brand-presentations-in-minutes-with-gemini-in-google-slides) that Gemini turns a description into an outline and, once you approve it, into a complete presentation. Gemini can adopt the style of an existing presentation and use documents or earlier presentations as sources. According to Google, the features are available to Workspace customers on the Business plans and on Enterprise Standard and Plus, and to subscribers of Google AI Pro or AI Ultra, and they launch initially in English only. For speakers, this means that the framework of slides for a conference talk now takes minutes, while the decision on a submission and on your appearance still rests on your topic, your abstract and your own cases.

**Key points at a glance**
- Google announced the "Generate presentation" feature in Google Slides on 23 July 2026 on the Google Workspace Blog, written by Dennis Hu, Group Product Manager for Google Slides.
- Gemini first produces an outline for you to review and builds the slides only after you approve it.
- An existing presentation can serve as the style template, and files such as Google Docs and PDFs as well as earlier presentations can serve as sources.
- The feature is available on Workspace Business and on Enterprise Standard and Plus, and with Google AI Pro or AI Ultra, initially in English only.
- In this article's worked example, six new talks a year at an assumed six hours saved per talk add up to 36 hours.

## AI presentation: what Google announced for Slides on 23 July 2026

The post was written by Dennis Hu, Group Product Manager for Google Slides, and it opens with a diagnosis: according to Google, building a great deck by hand takes a lot of time and design expertise. Google's answer is a workflow that starts at slides.new with the entry "Generate presentation", or in the Gemini side panel of a blank presentation. You describe what you need and add reference files. Google gives Google Docs and PDFs as well as previous decks as examples. If you select an existing deck, Gemini matches the style of your brand automatically, as the post puts it.

During creation, Gemini may ask about length and tone as well as the focus of the slides, and it suggests further relevant files from Google Drive. From all of this input, it produces a structured outline that you review before a single slide is built. Only after you approve does Gemini generate the deck in the background and notify you when it is done. Every slide stays editable. Through instructions, Gemini changes layouts and colour schemes, adjusts generated charts and rewrites text on individual slides to make it more professional or more concise.

Google states the availability in the same post: the feature is open to Workspace customers on the Business plans and on Enterprise Standard and Plus, and to subscribers of Google AI Pro or AI Ultra. All features launch initially in English only. The announcement contains no figures on time saved or on the quality of the results, and it says nothing about how uploaded files are handled.

## What an AI presentation is and what Gemini takes over

An AI presentation is a presentation in which a language model generates the outline and the design, including the slide text, from a description and source files, and a person then reviews and adjusts it. For a speaker at a professional conference, the work on a talk therefore splits into parts. According to the announcement, Gemini takes over the first ones, and the last ones stay with you.

**What Gemini in Google Slides takes over and what stays with you**
Task | Gemini according to Google | Speaker
Outline | proposes it from the description and source files | checks it against your own thesis and cuts what your audience does not need
Design | adopts the style of a selected presentation | decides which slide appears on stage at all
Slide content | writes text from the uploaded files | supplies cases with figures and your own experience, which have to be in those files
Revision | changes layout and charts as well as text on instruction | decides during the rehearsal which slide stays
Performance | the announcement gives no information | speaking and timing, plus the Q&A

## Slide frameworks are becoming cheap, and the talk is decided by your own cases

When every speaker can have a designed presentation within minutes, slides hardly set talks apart any more. What can set them apart is the material Gemini does not know because it is in none of the uploaded files: a project from your practice, the figure your client gave you, or the question someone asked you after your talk last year. We described how programme committees choose their talks in the article [What a programme committee looks for](/en/blog/what-a-programme-committee-looks-for/). Slides play a small part there, because a submission usually consists of a title and an abstract as well as the speaker bio.

This changes the question of whether you should have slides designed for you. The article [Having conference slides designed](/en/blog/have-conference-slides-designed/) covers what on a slide is design work and what is content. With Gemini in Google Slides, design comes within reach of every speaker, and the content remains the task you are booked for. How you use AI tools along the whole path from abstract to rehearsal is covered in [Using AI as a speaker](/en/blog/using-ai-as-a-speaker/).

## Calculation: what an AI slide framework saves across six talks a year

Google's announcement gives no time saving, so this section calculates with assumptions of our own, which you can replace with your figures. Suppose you develop six new talks a year and so far need eight hours per talk for the outline and design of the slides, including the wording. Suppose further that with Gemini, the draft and the design take two hours, including the review of the outline. You then save six hours per talk and 36 hours a year, which is four and a half working days of eight hours. The calculation is an example.

That time belongs in tasks that Gemini does not do. If an abstract for a submission costs you about two hours, 36 hours leave time for 18 additional submissions to conferences. You can also put the hours into rehearsals in front of an audience or into collecting cases with figures. Each of these tasks changes the result of your performance more than the layout of the slides.

## Limits of Gemini in Google Slides: language, plan and confidential data

Three points from the announcement and its gaps affect your planning. First, the feature launches in English only, and for conference talks in German the post gives no date for further languages. Second, it requires one of the named Workspace plans or a subscription to Google AI Pro or AI Ultra.

Third, you decide yourself which files you upload as sources. The announcement says nothing about how they are handled. Client cases with figures and names that a client shared with you in confidence therefore belong in a source file only after you have asked the client and your company's IT department. A speaker who anonymises anyway, for instance by sector and order of magnitude, can upload cases in that form.

## How to create an AI presentation for a conference talk in five steps

1. First write down the thesis and the key messages of your talk in your own words, before you describe anything to Gemini. The brief to the AI is then your own thought, and the outline you receive for review reflects it.
2. Put your cases with figures and sources into one document and add it as a reference file. Since Gemini also suggests additional files from Google Drive, check whether only files that belong in this talk are included.
3. Review the outline as if you were the programme committee. Every slide has to support one of your key messages. Slides that merely summarise material from the web, you cut.
4. Select as the style template an existing presentation in the design you need for the appearance. If you have a template from the organiser, it is also suitable as a style template in our reading of the feature.
5. Rehearse the talk aloud with a stopwatch, ideally in front of one person from your target audience. After the rehearsal, cut every slide you merely read from.

**Anyone who creates an AI presentation with Gemini in Google Slides gains time on the slide framework and does best to invest it in cases of their own with figures and in rehearsals, because the framework comes together equally fast for every speaker.**

**How Stage Starter gets you onto conference stages**
Stage Starter is a speaker agency for experts from B2B fields. We research the conferences that suit your topic and your ideal clients, and develop the talk with you. For every call for papers we write several versions of the abstract and submit in your name. We keep track of the deadlines and follow up with the organisers. Our founder Marin Niehues brings the experience of more than 100 invitations on six continents, and Sessionize named him Most Active Speaker in 2024 and 2025. In an [initial consultation](/en/erstgespraech/) we work out which stages suit you.

**On 23 July 2026, Google announced that Gemini in Google Slides creates an outline from a description and, after approval, a complete presentation, initially in English only and for certain plans. For speakers, the effort for the slide framework falls. A talk gains from cases of your own with evidenced figures and from a rehearsed performance, and the conference decides on a submission first by topic and abstract.**

## Frequently asked questions

### How can I create an AI presentation?

In Google Slides, you choose "Generate presentation" at slides.new or open Gemini in the side panel of a blank presentation and describe what you need. Gemini creates an outline for you to review and builds the slides after you approve it. According to Google, this works initially in English only.

### Which plans include Gemini in Google Slides?

According to Google, the feature is available to Workspace customers on the Business plans and on Enterprise Standard and Plus, and to subscribers of Google AI Pro or AI Ultra. The announcement of 23 July 2026 contains no price.

### Does Google's AI presentation work in German?

Not at launch, because Google writes that all features launch initially in English only. The post names no date for further languages.

### Should I let an AI create my slides as a speaker?

It is worth it for the framework: Gemini delivers the outline and the design within minutes, and you review both afterwards. The cases with figures and the experience for the conference talk have to come from you, and every slide should support one of your key messages.

### Do programme committees assess the slides in a call for papers?

A submission usually consists of a title and an abstract as well as the speaker bio. The programme committee selects on that basis. The slides mostly come after acceptance, so your abstract decides the submission first.