Introduction

Hello, I’m Mia Sato, AI Research Lead.

In this article, I will look at Google Flow from the perspective of how it may be used for prototyping promotional content, aligning stakeholders, and reducing rework in the production process.

Google Flow is a creative tool that allows users to generate AI videos based on text and images. With this latest update, features such as Gemini Omni Flash, Flow Agent, and Flow Tools have been added. This suggests that Flow is evolving beyond simply creating videos, and is beginning to support the broader creative process, including planning, revising, and organizing content.

At some GDX client companies, conversations like the following often take place:

“The direction of this promotional video looks good, but it may be a little too long for social media.”

“We also want the tone to match the materials used for the landing page.”

“Then let’s first organize the assets and recreate them for each use case.”

Even when creative assets already exist, teams often need to remake them again and again to fit different sizes, durations, and media channels. These small rounds of rework can easily accumulate in EC and promotional operations.

That is why a tool that can handle not only video generation, but also revisions and the creation of multiple variations within the same workflow, may help reduce this friction.

In this article, we will review the latest updates to Google Flow from the perspective of reducing rework time in promotional production and EC operations.

What Has Changed with Google Flow?

Google Flow is an AI video creation tool from Google Labs. It allows users to create video clips and scenes based on text and images.

Until now, Flow has mainly been used by entering an idea, generating a video, selecting the best result, and then editing it.

What stands out in this latest update is that Flow seems to be moving from “a place to create videos” toward “a place to create together with AI.”

The three most important changes are as follows.

1. Video Creation Is Moving from One-Shot Generation to a More Conversational Workflow

The key update this time is the introduction of Gemini Omni Flash into Flow.

Omni Flash is a model that can combine text, images, video, and audio to create and revise videos. In Flow, this makes it easier to revise visual content through natural language while using both real footage and AI-generated materials.

For example, a team might first create a product introduction video.

After that, they could revise it through conversation by giving instructions such as:

“Make the opening more dynamic.”

“Shorten this for social media.”

“Keep the same person, but place them in a different scene.”

This is the kind of production process Flow is moving toward.

AI video generation has often felt like a process of generating again and again until a good result appears. What is especially promising about the updated Flow is that it aims to connect the creation process with the revision process after the video has been generated.

2. Flow Agent Makes It Easier to Ask for Support from Ideation to Asset Organization

Another major update is Google Flow Agent.

Flow Agent is a creative agent that supports tasks such as planning, story development, dialogue writing, generating multiple ideas, batch editing, and organizing creative assets.

In video production, the difficult part is not only the generation itself.

Teams also need to decide:

Which message should the video focus on?

Who is the target audience?

In what order should the information be presented?

How should the created assets be organized?

These tasks before and after production often take a lot of time.

Flow Agent is designed to support this part of the workflow. Rather than simply outputting videos, the AI works alongside the production process and helps move it forward.

From a practical business perspective, this could be especially useful. For example, even when using the same campaign assets, the way they are presented may differ slightly across an EC site, social media ads, and in-store digital signage. If promotional, creative, and product teams need to review these adjustments every time, even having Flow Agent organize channel-specific ideas in advance could help reduce rework.

3. Flow Tools Can Turn Repeated Tasks into Small Custom Tools

Another feature worth noting is Google Flow Tools.

This feature allows users to create their own creative tools and workflows using natural language. For example, Google explains that users may be able to create tools for video resizing, image editing, or custom shaders without writing code.

When applied to EC operations, the potential use cases are easy to imagine.

Convert videos into vertical format for social media every time.

Create images with a consistent tone for landing pages.

Standardize video length and structure for each campaign.

These repeated tasks may be turned into small tools that fit each team’s own production workflow.

Rather than expecting perfect automation from the beginning, it may be more practical to think of this as a way to make “the same revisions we request every time” a little easier. Seen this way, the feature feels highly relevant to real-world operations.

GDX’s Perspective: Where Can This AI News Be Applied in Business?

From the perspective of EC operations, the latest Google Flow update seems useful not only for video production itself, but also for recreating promotional assets, reviewing materials, and developing multiple variations.

Task: Social Media Ads and Short-Form Video Production

Purpose:
Quickly create multiple versions from a single key message for platforms such as Instagram, YouTube Shorts, and TikTok.

Why it works:
Flow Agent can generate multiple ideas, while Omni Flash can help teams adjust video length and presentation style through conversation.

Evaluation point:
For products with strict requirements around human representation or brand expression, teams should proceed carefully. It may be better to start with product usage scenes or campaign announcements.

Task: Video Drafts for Landing Pages and Feature Pages

Purpose:
Quickly create short videos for first-view sections or product explanation cuts.

Why it works:
Flow is moving toward a workflow that connects images and videos within the broader creative process. This makes it suitable for creating initial drafts.

Evaluation point:
Rather than publishing the output as-is, it may be more effective to use it for review sessions with designers, merchandisers, or product teams.

Task: Revising Product Introduction Videos

Purpose:
Revise only part of a video that has already been created, reducing the need to remake it from scratch.

Why it works:
Because Flow is moving toward conversational editing, it may become easier to adjust elements such as subtitles, background music, the opening sequence, and scene length.

Evaluation point:
Information such as prices, campaign periods, and inventory availability can have a major impact if incorrect, so these details should always be reviewed by humans.

Things to Consider Before Implementation: From Information-Organizing AI to Flow Prototyping and Stakeholder Alignment

When introducing Google Flow into promotional production, it is important not only to think about which Flow features to use, but also to consider how multiple AI tools can be combined.

In practice, rather than expecting Flow to solve everything on its own, it is more effective to treat the process as one connected workflow: organizing information before using Flow, prototyping with Flow, and then aligning with stakeholders. Each step can be supported by different AI tools.

1. First, Use an Information-Organizing AI to Align the Basics

AI video tools are strongly affected by the quality of the materials and background information provided as input.

That is why, before using Flow, teams should at least organize the following information:

Product images

Key product messages

Target users

Intended media channels

Video length

Expressions to avoid

Brand tone

However, all of this does not need to be organized manually by humans. In fact, the information preparation stage before using Flow is exactly where another AI tool can provide support.

For example, by having AI read product reviews, landing pages, past ad copy, social media posts, and meeting notes, teams can more easily organize key messages, target users, brand tone, expressions to avoid, and possible video structures.

Among the tools introduced previously, NotebookLM is well suited for organizing information based on multiple sources such as product materials, landing pages, and reviews.

By letting AI handle this initial organization instead of doing everything manually, the instructions given to Flow can become much more specific.

2. Next, Use Flow to Create Video Prototypes

Once the basic information has been organized, the next step is to create video prototypes with Google Flow.

The important point here is not to aim for a finished version from the beginning. Flow is likely to show more value when used to create promotional video drafts or multiple ideas, rather than as a final production tool right before publication.

For example, teams could create multiple short-form video ideas for social media ads, test different versions for the first view of a landing page, or change only the opening of a product introduction video to compare how the message is perceived.

In this way, Flow becomes easier to incorporate into real operations when it is used not as “a tool that creates the final product in one shot,” but as “a tool that quickly creates prototypes for stakeholder alignment.”

3. Finally, Align with Stakeholders

After prototyping videos with Flow, the next step is stakeholder alignment.

In promotional production, teams need to review not only the quality of the video itself, but also the direction of the message, the brand tone, the way the product is presented, and whether the content is suitable for each media channel.

For example, the promotional team may review the messaging. Designers may check the tone and visual appearance. Merchandisers or product teams may review the accuracy of the product representation. Managers may decide whether the content can be published.

AI can also support this stage by organizing review points or extracting revision tasks from meeting notes. If the outcomes of stakeholder discussions can be converted into the next set of revision instructions for Flow, the revision process can become smoother as well.

Summary

The latest update to Google Flow is not simply an improvement in video generation.

With Gemini Omni Flash, video creation is moving away from one-shot generation and closer to a process where users can revise videos through conversation. With Flow Agent, it becomes easier to handle planning, asset organization, and the creation of multiple ideas. With Flow Tools, there is also potential to turn repeated tasks into small tools tailored to each team’s workflow.

From the perspective of EC operations, Flow seems well suited for SNS ads, landing page assets, product introduction videos, and multiple versions of campaign materials. However, rather than handing over all final production work to Flow from the beginning, it may be more practical to use it for drafts, comparison ideas, and early review sessions.

It is also important not to rely on Flow alone. Before using it, teams should use an information-organizing AI to organize product materials, landing pages, reviews, past ads, and meeting notes, and to align key messages and brand tone. Then, by using Flow to create video prototypes and reviewing them with stakeholders, teams may be able to reduce rework across the entire promotional production process.

What matters is not only being able to create videos, but also aligning the assumptions before creating them.

What matters is not only increasing the number of assets, but also making the review and revision process smoother.

The latest Google Flow update is a practical development worth watching for anyone thinking about how to combine AI tools in promotional production.

References

Official reference: New agents, mobile apps and Gemini Omni for Google Flow and Google Flow Music / Google Blog / https://blog.google/innovation-and-ai/models-and-research/google-labs/flow-updates/

Official reference: Introducing Gemini Omni / Google Blog / https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-omni/

Official reference: Get started with Google Flow / Google Flow Help / https://support.google.com/flow/answer/16353333

Commentary / Expert reference: Google Makes It Easy to Deepfake Yourself / Reece Rogers / WIRED / https://www.wired.com/story/google-makes-it-easy-to-make-a-deepfake-of-yourself/

Commentary / Expert reference: Google Flow AI video editing & music tools getting dedicated apps and Omni upgrades / Ben Schoon / 9to5Google / https://9to5google.com/2026/05/19/google-flow-video-music-ai-apps/

For more information about GDX Inc., please visit the link below.
Company website: https://gdx.inc/

Part of this article was created with the support of ChatGPT and then edited and revised by the author. The content reflects the author’s personal views and does not represent an official statement or position of GDX Inc. The information is provided for reference purposes only. Please refer to official announcements and primary sources for confirmation.