Kanata Studio BetaNEW

ImageStudio:A generative AI image creation studio

ImageStudio creates images such as announcement banners and training notices from your instructions and Design System. Specify reference images and internal documents, create one to four options, and save only the one you choose to Media.

Reviewing exhibition announcement banner options created in ImageStudio, along with the settings used to create them

ImageStudio overview

ImageStudio is an image creation feature connected to Kanata's Library and Design System. Unlike simply giving instructions in a chat, you specify the number of images, reference images, reference documents, Design System, size and quality on a dedicated screen in addition to your instructions, and create while comparing options. Saved images are registered in Media and managed with folder permissions, so images are less likely to end up scattered across personal PCs and separate tools.

01

What it covers

Creating images such as announcement banners, notices for training and internal events, and diagrams for documents

02

Best suited for

Organizations that outsource image creation or rely on a few people, and want each department to create images while following brand and generative AI usage rules

03

What you can aim for

Reducing back-and-forth on image requests and revisions, and making it possible to check later who created what, with which references and which Design System

Problem

Challenges organizations face in creating images

Surveys show that while infringement of rights such as copyright and leaks of internal information are cited as risks of generative AI, only some companies have risk management in place for risks specific to AI.

Share citing "may infringe copyright and other rights" as a risk of generative AI (Japan)

31.3%

The most common answer was "security risks such as leaks of internal information" at 46.4%, followed by "problems with the accuracy of outputs" at 35.3%.

Source: Ministry of Internal Affairs and Communications, "White Paper on Information and Communications in Japan 2026", Figure I-2-1-8 (survey conducted January–February 2026; in Japan, 515 managers or above at companies)

Share of companies citing "legal regulations such as copyright and privacy protection" as a concern or issue with generative AI

18.1%

Up to three answers allowed; base of 10,312 valid responses. "Risk of information leaks" was 33.5%.

Source: Teikoku Databank, "Survey on Corporate Trends Regarding Generative AI (March 2026)", Figure 5 (survey conducted March 17–31, 2026; 10,312 valid responses)

Share of companies implementing risk management specific to AI and generative AI

14.2%

The question was asked of companies that answered they "have adopted", "are trialing" or "are considering" AI.

Source: IPA, "DX Trends 2026", Figure 3-18 (survey conducted April 17–June 12, 2026; 1,799 responses)

Challenges on the ground

We want to create images with generative AI, but concerns about rights and expression risks make it "hard to set usage rules"

Each person uses a different tool, so "brand colors and logo usage are not consistent"

Image creation is outsourced or left to a few people, and "revisions take a long time"

Images end up scattered across personal PCs and tools, and "no one knows where they are"

Simply adopting separate image generation tools can leave reference documents, created images and records of generative AI use split across tools. The creation process and its records need to sit within your business platform.

Scenes

How ImageStudio is used

Examples of how department staff use ImageStudio. The video shows how to create an image announcing a training course.

The companies and people shown are fictitious.

Scene 01

An image announcing a training course

HR and admin staff

BeforeEvery time they requested an image announcing a training course, the back-and-forth on revisions took a long time.

  1. Specify instructions and reference imagesIn addition to instructions, specify reference images ("material" or "mood") and internal documents such as the training notice text.
  2. Compare optionsCompare options where the first follows the instructions and the rest each change one of composition, background or style, then pick one.
  3. Review and saveReview the content, then save the chosen option to Media. Saved images are used in Library documents and elsewhere.

With KanataStaff compare options themselves and announce the training with the image they have reviewed.

Challenge solved: revisions take a long time

Scene 02

A trade show announcement banner

Sales and marketing staff

BeforeEach person used a different tool, so brand colors and logo usage were not consistent.

  1. Choosing a Design SystemChoose one Design System when creating, and settings such as colors and the brand promise are included in the instructions for the image AI.
  2. Placing the logoThe logo is not drawn by the image AI; the logo image registered in Design System is overlaid after creation.
  3. Creating other sizesFrom the chosen option, create the other sizes the announcement needs. The current options remain as they are.

With KanataWhoever creates them, announcement images follow the same color and logo rules.

Challenge solved: brand colors and logo usage are not consistent

Scene 03

A social media image for an announcement

Admin and PR staff

BeforeImages ended up scattered across personal PCs and tools, and no one knew where they were.

  1. Refer to the draftCreate the image with the announcement draft in the Library specified as a reference document.
  2. Save to MediaOptions do not go into the Library until saved, and only the chosen option is saved to Media. Who can view it is managed with folder permissions.
  3. Check the generation informationSaved images show the source "ImageStudio" in the Library.

With KanataImages are stored together in Media, and people review the content before using it externally, using the displayed source as a guide.

Challenges solved: no one knows where they are, hard to set usage rules

These scenes are examples of use. The people shown are fictitious.

Platform

The advantages of being part of an AI platform for organizations

ImageStudio runs on the same foundation as Kanata's Library, Design System and AI Agent. Referring to the documents needed for image creation, managing where images are saved and who can access them, and recording usage all happen on a single foundation.

Referring to Library documents and images

Choose reference images from Media and reference documents from the Library.

Managed with folder permissions

Saved images are registered in Media, and folder permissions determine who can view them.

Choosing a Design System

Choose one Design System for each creation, the same Design System used by SlideStudio.

Requests from the AI agent

AI agents permitted through execution permissions create images in response to requests in chat. Depending on the autonomy level, the requesting user reviews the plan before creation.

Use in the Library

Select saved images from Media and use them in Library documents and elsewhere.

Managing consumption with credits

Image creation is managed with the same credits as other features. The estimated consumption against your available credits is shown before you create.

Features

ImageStudio features

In ImageStudio, you create images in the order "input → create → review → save". Work in progress is managed inside Studio, and only the option you choose to save is registered in Media.

Instructions and reference images

Create by specifying instructions, reference images ("material" or "mood") and reference documents.

Compare options and regenerate

Compare up to four options and regenerate from the one you choose.

Brand and records

Applies your Design System and shows the source of saved images.

Main features

Decide

Specify instructions and reference images

Specify the image name, instructions, reference images, reference documents, size and quality

Specify up to four reference images, and for each choose "material" (use what is shown as it is) or "mood" (use only the colors and feel as reference). Choose the size from square, landscape, wide, portrait or custom, and the quality from rough, standard or fine; the estimated consumption based on quality and number of images is shown before you create.

The ImageStudio input screen, checking the instructions, number of images, reference images and estimated consumption
Review

Compare options and regenerate

Compare an option that follows the instructions with options that vary composition, background and style

When you create two or more images, the first follows the instructions and each of the rest changes one aspect not specified in the instructions. "What AI decided" shows what AI decided for parts not covered by the instructions. Regenerating from a chosen option uses the settings that option was created with, and the current options remain.

The ImageStudio review screen, choosing one of four options and saving or recreating it
Brand

Design System and logo examples

Apply colors, logo and brand promise

Choose one Design System, and settings such as colors and the brand promise (tone, prohibited items) are included in the creation instructions; you also choose whether to add the logo. People check before saving that the output follows the settings. The logo is not drawn by the image AI; the registered logo image is overlaid after creation.

The logo settings of the "MINATO" design system, checking the logo image and live preview
Save

Saving to Media and generation information

Save only the chosen option and record AI creation

Options remain in Studio as drafts and do not go into the Library until saved. Saved images are registered in Media and show the source "ImageStudio" in the Library list and details.

An image created and saved in ImageStudio, recorded in the Library along with its source

Column

Related columns

FAQ

FAQ

QHow can we tell an image was created with AI?

Saved images show the source "ImageStudio" in the Library list and details. Before using images externally, please have a person review the content according to your internal generative AI usage rules.

QCan we add text to images?

Yes. Include text such as headlines and captions in the instructions, and it is generated together with the image. Text may come out distorted, so please check it before saving.

QWhich formats can be used as reference images?

PNG, JPEG and WebP. You can specify up to four reference images and choose "material" or "mood" for each.

QWhich image sizes can be created?

Square (1024×1024), landscape (1200×630), wide (1920×1080), portrait (1080×1350), and custom, where you enter the width and height. A chosen option can also be created in another size.

QDoes creation consume credits?

Yes. Consumption depends on quality (rough, standard, fine) and the number of images. The estimated consumption against your available credits is shown before you create.

QCan we ask the AI agent to create images?

Yes. If you allow ImageStudio in the AI agent's execution permissions, it creates images in response to requests in chat. Depending on the autonomy level, the requesting user reviews the plan before creation. A record of the creation remains in the ImageStudio screen.

Last updated: