M365 Copilot — Visual Creator Agent

Licence required before you start

Visual Creator requires the Microsoft 365 Copilot paid add-on — $30/user/month (Enterprise) or $21/user/month for organisations with up to 300 users (Business). It is not available on standard M365 plans or free Copilot Chat. Image generation is subject to Microsoft's content policies — prompts that request real people's faces, branded logos, or inappropriate content will be declined. Verify under Billing → Licences in the M365 admin centre. Pricing reflects standard US list pricing and may vary by region, currency, or enterprise agreement.

What's new — June 2026

Copilot Cowork is now generally available and billed on usage (Copilot Credits, $0.01/credit pay-as-you-go); the Work IQ APIs also reached GA. Visual Creator itself is a built-in Copilot agent included under your Copilot licence and is not separately metered — but if your organisation builds custom agents that ground in M365 data via Work IQ, that usage is metered in Copilot Credits. See the licensing/comparison guide for the full cost picture.

June 17, 2026 — June 16 GA wave: Cowork GA + Copilot Credits usage billing + Work IQ APIs GA.

At a glance

What it is
A visual content agent. Describe what you want
in plain text — it generates an image.

Key use case
Presentation visuals, concept illustrations,
internal comms banners, training materials.

What it cannot do
Generate images of real people by name.
Reproduce copyrighted logos or brand marks.

Use case	What Visual Creator produces	For
Presentation visuals	Illustrative images for slide decks and reports	PM · Marketing · Comms · EA
Concept diagrams	Process flows, abstract concept illustrations	Consultant · Designer · L&D
Internal banner images	Intranet headers, Teams channel banners, event visuals	Comms · HR · Admin
Training materials	Scenario illustrations, role-play visuals	L&D · HR · Manager
Report covers	Professional cover visuals for executive reports	Analyst · Finance · Strategy

How to use Visual Creator

Open Copilot and select Visual Creator

Go to m365.cloud.microsoft. In the left panel under Agents, select Visual Creator. It is built-in — no manual installation needed.

Write a descriptive prompt

Type a description of the image you want. The more specific, the better: include style ("professional, flat illustration"), colours, mood, and subject matter. Avoid mentioning real people, logos, or brands.

Review the generated images

Visual Creator generates several variations. Review them and select the one closest to what you need. If none work, refine your prompt and regenerate.

Refine with follow-up prompts

Ask for changes in natural language: "Make it lighter", "Remove the background", "Add a cityscape in the background", "Make it more abstract". Iterate until the output fits your purpose.

Download and use

Download the generated image as PNG. Insert into PowerPoint, Word, Teams, or your intranet. Generated images are subject to Microsoft's content policy — review usage rights at microsoft.com/en-us/designer before using externally or in client-facing materials.

Writing better prompts — what to include

The four elements of a strong visual prompt

Works for any type of image

Start here

Include in your prompt

Subject — what is in the image (people, objects, scene, concept)
Style — photorealistic, flat illustration, watercolour, 3D render, minimalist
Mood — professional, warm, futuristic, calm, energetic
Colours — "blue and white palette", "muted earth tones", "dark background"

Example prompts

A flat illustration of a team collaborating around a digital screen, blue and white palette, professional, no faces visible.
Abstract concept image of data flowing through a city, dark background, electric blue accents, futuristic style.
Minimalist banner image of an industrial plant at sunrise, warm tones, photorealistic, wide format.
Professional cover illustration for an executive report on AI adoption, geometric shapes, navy and gold, clean and modern.

Avoid vague prompts like "a good image for my presentation". Specific prompts produce usable outputs. Vague prompts produce generic stock-photo results.

Use cases — by role and scenario

What to ask and what to expect

Executive presentation visuals

Strategy · Finance · PM · EA

High value for deck preparation

Common scenarios

Title slide illustration for a strategic report
Section divider visuals in a board presentation
Concept image for a transformation initiative
Cover image for a quarterly business review

Example prompts

A professional cover illustration for a digital transformation strategy deck, dark blue and gold, geometric, clean.
An abstract image representing growth and change, upward motion, warm tones, minimalist style, no text.
A futuristic industrial facility at night with digital overlays, photorealistic, wide format, for a board presentation.

For brand-aligned visuals, specify your organisation's colours in the prompt. Visual Creator does not have access to your brand guidelines automatically.

Internal communications and HR

Comms · HR · L&D · Admin

Quick wins for internal content

Common scenarios

Intranet or Teams channel banner image
Training module illustration
Internal event poster or announcement visual
Manager communication newsletter header

Example prompts

A cheerful banner image for an internal wellbeing campaign, people in a bright modern office, flat illustration style, no faces, diverse group.
A training module header image about cybersecurity awareness, padlock and digital network, blue tones, clean and professional.
A wide banner for an internal AI skills programme, abstract brain and circuit motifs, purple and white, modern.

Requesting "no faces" or "silhouettes only" avoids content policy rejections and makes images more versatile for diverse workforces.

Known limitations and considerations

The prompt is declined for policy reasons. Visual Creator will not generate images of identifiable real people by name, reproduce brand logos, or produce content that violates Microsoft's content policy. Rewrite the prompt to describe generic subjects — "a professional in a modern office" instead of "a photo of [person's name]".

The output does not match the description. Very abstract or complex scenes are harder to generate accurately than simple, concrete descriptions. Break complex scenes into simpler components. "A clean desk with a laptop and a plant, natural light, minimalist" produces better results than "the perfect work environment".

The image looks generic or stock-photo-like. Add style qualifiers to push away from default outputs: "not a stock photo", "illustration style, not photorealistic", "flat vector art, no shadows". These push the output toward more distinctive visuals.

Colours are wrong or inconsistent. Specify colours explicitly using colour names or hex-like descriptions: "dark navy background, white foreground elements, gold accents". Avoid vague instructions like "corporate colours" — Visual Creator does not know your brand palette.