Canva AI: Every Current Feature Explained

AI Summary

Canva AI is the collective name for the artificial intelligence features built into Canva’s design editor. It covers conversational design, image and video generation, copywriting, layout building, spreadsheet analysis, interactive builds through Canva Code, background editing, resizing and translation. In the UK, access is included with whichever plan you hold: Free at £0, Pro at £100 a year for one person, Business at £180 a year per person, and Enterprise on application.

Every plan draws on a shared monthly AI allowance split into Standard, Premium and Ultra tiers, which replaced the older credit system entirely. The largest recent change is Canva AI 2.0, launched in April 2026, which added conversational and agentic design and is still reaching accounts in stages rather than being switched on for everyone at once.

Verified against Canva’s official pricing page, Help Centre and AI Product Terms on 17 September 2026.

Canva AI has changed more in the past twelve months than in the three years before it, and most guides still in circulation describe a version of the product that no longer exists. If you have read that Canva gives you 50 lifetime image credits, or 500 credits a month, or that the free plan covers 200 AI uses, all three are out of date. The credit system is gone. The allowance figures were restructured again this year. Several tools were renamed, and a whole conversational layer was added on top.

This guide sets out what Canva AI includes today under its current names, what the plans cost in GBP, how the monthly allowance is actually spent, which models sit underneath the output, and where the platform stops being the right tool for the job. For teams who want these tools embedded into an existing content process rather than used ad hoc, ProfileTree runs AI training built around the tools a business already has rather than around a generic curriculum.

What is Canva AI, and what does it actually include?

Canva AI is not one product. It is a set of features spread across the editor, each doing a specific job, and knowing which tool does what saves real time because the naming is inconsistent and several tools overlap.

The chat layer sits at the top. Canva AI assistant takes a description from the homepage or the editor and generates or amends a design in response. Below that, the tools group roughly by the kind of work they do.

Generating visuals

• Image generation with Canva AI produces images from text prompts, with a quality tier you choose at the point of generation.
• Create a video clip with Canva AI turns a prompt into a short clip with synchronised audio, and Image to Video animates a still.
• Generate elements with Canva AI creates individual shapes, 3D objects, charts and forms inside an existing layout rather than a whole image.
• Magic Layers turns generated output into properly editable layouts, so an asset can be pulled apart element by element instead of arriving flat.

Editing what you already have

• Magic Edit replaces a selected part of an image from a prompt. Magic Eraser removes objects and fills the space. Magic Expand extends the canvas past the original edges.
• Magic Grab, Magic Morph and the background tools handle object extraction, text and shape effects, and background replacement.
• Image Upscaler, Mockups, Beat Sync and Highlights cover the smaller production jobs.

Writing and adapting

• Magic Write drafts captions, headings, body copy and outlines inside Canva Docs and on design canvases.
• Magic Resize and Translate reformat an existing design for different channels and languages. For teams working across formats, this pairing still returns the most time per unit of effort in the whole set.
• AI Voice generates narration from a script.

Working with data and interactivity

• Magic Charts, Magic Insights and Magic Formulas work inside Canva Sheets, turning raw data into visualisations, summaries and formulas.
• Canva Code builds interactive elements such as forms, calculators and widgets from a written description.
• Canva Grow handles AI ad creation and ad insights.

One naming point worth settling. Canva now leads with “Canva AI” as the umbrella term, and its own product navigation is organised that way, but “Magic Studio” has not disappeared from Canva’s Help Centre and several individual tools kept the Magic prefix. Both names are still in circulation, and they refer to broadly the same set of features.

Ciaran Connolly, founder of Belfast digital agency ProfileTree, puts it plainly: “The value for most SMEs isn’t in any single tool. It’s in having a connected workflow where one piece of content can be quickly adapted across every channel the business uses. That’s where Canva AI saves real time.”

What changed with Canva AI 2.0?

Canva AI 2.0, announced in April 2026, is the largest change to the product since launch, and it shifts Canva from a design tool with AI features to a conversational platform where the design is assembled through dialogue.

Canva describes four capabilities underneath it. Conversational design starts creation with a description rather than a template, and keeps context as the idea develops. Agentic orchestration gives the assistant access to Canva’s design engine so it can select and coordinate tools towards a stated goal. Layered object intelligence builds output from individual editable objects, so changing one element does not disturb the rest. The Memory Library carries your style and brand across projects so suggestions improve with use.

Six workflows sit on top of that foundation:

• Connectors link Canva AI to tools a team already uses, including Slack, Gmail, Google Drive, Google Calendar, Notion, Zoom, HubSpot, Microsoft, Atlassian and Linear.
• Scheduling runs tasks in the background on a set cadence, so a weekly batch of social content is ready when you log in.
• Web research gathers and structures information from the web directly into a design as editable content.
• Brand Intelligence applies your fonts, colours and style from the first output, and can reapply an updated brand across existing work.
• Canva Code 2.0 builds interactive experiences from a single prompt and now imports HTML files for editing in the visual editor.
• Sheets AI generates structured, populated spreadsheets from a description.

Here is the part most coverage leaves out, and it matters more than the feature list. Canva AI 2.0 launched as a research preview and has been rolled out in stages. In August 2026, Canva cut its expected revenue growth rate from 30% to 20%, with chief executive Melanie Perkins telling Fortune that demand for the new AI features significantly exceeded expectations and that the company chose to slow the rollout while it rebuilt the architecture and reduced the cost of completing an AI task. A successor version has since begun reaching accounts on the same staged basis.

The practical consequence for a UK business: what Canva markets and what your account can do this week may not match. Check the editor before planning a process around a named feature.

Which AI models sit behind Canva AI?

Canva runs a mixed stack: one design model it built itself, plus partner models for specific jobs. This explains why some output arrives as separate editable layers and other output arrives flat.

The centrepiece is the Canva Design Model, which Canva describes as a foundation model built to understand the structure, hierarchy and complexity of real-world design rather than flat images. That training is the reason generated output can arrive as individual objects you can edit one at a time.

Image generation draws on Leonardo.ai, the generative platform Canva acquired in 2024. Video clip generation runs on Google’s Veo 3, producing a 16:9 clip of up to eight seconds per prompt with synchronised audio, including dialogue and sound design.

Canva’s AI Product Terms describe third-party providers as Technology Partners and state that your input may be shared with them for the specific purpose of providing the functionality. Perkins has also said publicly that Canva had been relying too heavily on frontier models from external providers and has been moving more of the work onto its own models to bring the cost per task down.

Two smaller points that matter for anyone producing client work. Canva applies provenance metadata to AI output, and its terms prohibit removing or altering those tags. The same terms prohibit presenting AI-generated content as human-generated. If your business publishes under a disclosure policy, both of these are worth writing into it.

How much does Canva AI cost in the UK?

There is no separate charge for Canva AI. It is included with whichever plan you hold, and the plan sets your monthly allowance. The figures below are Canva’s UK annual prices, taken from its UK pricing page. Monthly billing is available on every paid tier at a higher effective rate, and UK subscriptions attract VAT on top.

The free tier remains genuinely usable for occasional work, though its AI allowance is now tight enough that regular generation will exhaust it quickly. For a business producing content weekly, Pro at £100 a year works out at a little over £8 a month before VAT, which holds up well against standalone AI image tools.

The step from Pro to Business is the one that catches people out. Pro is a single-person licence. The moment a second person needs their own login, you move to Business at £180 per person per year with no volume discount before Enterprise, so a three-person team is £540 a year. Brand Controls, approval workflows and AI admin controls only appear at Business level, and for teams producing content under one brand identity, those governance features are usually the real reason to upgrade rather than the larger allowance.

One detail to budget around: allowances are per person and are not pooled across a team or organisation. A heavy user on your team cannot borrow from a colleague’s unused allowance.

How does the Canva AI allowance actually work?

Canva no longer uses credits. Every AI tool draws on a single shared monthly allowance, and different tiers of tool draw it down at very different rates. This is the part most guides still describe incorrectly, and the figures have been restructured more than once.

Because it is one shared pool, the published numbers are alternatives rather than separate budgets. A Pro account gets up to 200 Premium uses or up to 20 Ultra uses, not both. Canva also notes that its own figures assume simple prompts and the lighter tools within each tier, so real usage typically lands below the headline number. Generating and iterating on a forty-slide deck through conversation will consume far more allowance than a single text-to-image prompt.

Four details worth knowing before you plan around it, all confirmed on Canva’s Help Centre page on AI usage:

1. Some tools sit outside the allowance completely. Magic Write, Translate, photo and video animations and AI Voice are covered by your plan under fair use rather than metered.
2. Hitting the limit is not a hard stop. Standard and Premium tools continue with short pauses between generations. Ultra tools stop until your allowance resets.
3. Resets follow your billing date, not the calendar. Paid plans reset monthly on the billing date, including annual subscriptions. Free accounts reset at midnight UTC on the first of the month.
4. Trials are not representative. Pro and Business free trials are capped at twenty Standard or Premium uses with no Ultra access at all, which is the same as the Free plan. If you are evaluating Canva on a trial, you are not seeing the allowance you would pay for.

Two routes exist for more capacity. AI Pass is a recurring monthly add-on for Pro and Business subscribers. AI top-ups are one-off purchases on Business, valid for thirty days and tracked separately from the base allowance.

There is also a quieter drain to watch. If you connect Canva to an external AI assistant to generate designs from another chat interface, those generations come out of the same Canva allowance. Third-party apps from Canva’s Apps Marketplace work the other way around and are not included, though they may carry their own limits or subscriptions.

A caveat on all of the above. Canva’s AI Product Terms describe usage limits as operational controls rather than fixed entitlements, and reserve the right to change them for performance, fairness, security or reliability reasons. Treat the numbers as current rather than contractual, which is also why so many published guides contradict each other.

How do you get better output from Canva AI?

Output quality tracks prompt specificity more closely than it tracks the model. “A marketing banner for a Belfast accountancy firm” produces generic results. “A clean, professional flat-lay photo of a desk with a calculator, coffee and financial documents, natural light, white background, wide format” produces something usable.

A working sequence for an SME content team looks like this:

1. Draft the words first. Use Magic Write inside Canva Docs to get a first version of the post, description or outline, then edit it to your brand voice before it goes near a design. Magic Write does not count against your allowance on paid plans, so this step is effectively free.
2. Generate imagery deliberately. Use specific, descriptive prompts and produce several options before committing. Choose the Standard quality tier for routine work and save the higher tiers for assets carrying a campaign.
3. Build the primary asset. Assemble the main design from templates and your generated images. Where the AI has produced a full layout, run Magic Layers over it so the elements stay editable.
4. Adapt with Magic Resize. Generate channel-specific variants from the primary design rather than rebuilding each format from scratch.
5. Review before publishing. Check for image artefacts, copy that does not read naturally, and brand inconsistencies.

A few prompting habits that consistently help: name the format and context first, include a style reference such as “flat lay” or “isometric graphic” to give the model a register to work within, use location-specific references when the content is for a local audience, and use Magic Edit to fix one problematic element rather than regenerating the whole image.

The thing to be clear-eyed about is what the AI will not do. It generates, it does not edit strategically. It will not tell you your headline is weak or that your copy misses your brand voice. That judgement still sits with a person, and it is the single most common difference between businesses getting value from these tools and businesses producing a lot of content that looks like everyone else’s. The same principle applies wherever AI sits inside a marketing process, which is why putting AI to work across marketing is a workflow question before it is a tooling question.

Can you turn Canva AI off, and does Canva train on your designs?

Both questions come up constantly from businesses in regulated sectors, and both have practical answers rather than policy answers.

Turning it off. On Business, Enterprise and Education plans, administrators can enable or disable AI tools from the admin controls. That switch is account-wide rather than per-user, which suits organisations needing a documented position on AI use. Free and Pro have no equivalent admin control, so the practical approach on those tiers is simply not to invoke the tools.

Training on your content. Canva’s AI Product Terms state that your privacy settings control whether Canva and its technology partners may use your data to improve AI services. Those settings are account-level and have been revised more than once as the product developed, so check what is currently enabled on your own account rather than relying on what was true last year or on what a guide says.

For businesses working under UK GDPR obligations, the sensible position is to document which tools are permitted, what categories of data may be entered into a prompt, and who signs off before AI-generated material reaches a client. Most of the risk in practice comes from an untrained team member pasting client data into a prompt, not from the platform itself. The connectors added in Canva AI 2.0 sharpen this point considerably, because an assistant with access to your email, calendar and file storage has a much wider reach than one working from a text box.

Can you use Canva AI images commercially in the UK?

In most cases yes, with caveats that Canva states plainly and most competing guides skip over.

Ownership. Canva’s AI Product Terms state that, as between you and Canva, you retain ownership of your input and you own your output. There is an important exception: output that modifies or incorporates Canva’s licensed library content is covered by the licence you hold for that content rather than owned outright. Edit a library photo with an AI tool and you do not own the result. Generate a presentation containing library photos and you own the presentation as a whole but not the individual images inside it.

Exclusivity. Canva states directly that outputs may not be unique and that other users may receive similar outputs. Your ownership does not extend to anyone else’s output. Canva also does not guarantee that output is cleared for use, and places responsibility for checking commercial suitability on you.

Standalone sale. You cannot sell AI-generated content on its own. It has to form part of an original design, which matters for anyone producing print-on-demand or template products.

The UK legal position. Works generated autonomously by AI sit awkwardly within the Copyright, Designs and Patents Act 1988, which was drafted around human authorship. Works where a person has made creative choices during generation, including through prompt writing, sit in a greyer area. For social posts, marketing materials and website imagery this rarely creates practical problems. For publication, broadcast or high-stakes commercial licensing, taking specific legal advice before relying exclusively on AI-generated images is a reasonable precaution.

Indemnity. Canva Shield covers trust, safety and privacy provisions across all plans. The AI output indemnification element, which is the part that actually reduces legal exposure, applies to Enterprise customers with a minimum of 100 seats. It is not included with Pro or Business.

Where does Canva AI stop being the right tool?

Canva AI’s advantage is not image quality in isolation. It is integration, because the AI tools sit inside a full design environment alongside templates, stock assets, publishing tools and brand management. For most SME teams that matters more than the technical ceiling of any single generation model.

The platform works well for:

• Teams without dedicated designers who need usable output quickly
• Multi-format content production, where Magic Resize saves significant rework
• Social media and marketing content at regular production volume
• Presentations, training materials and internal communications
• Spreadsheet summaries, charts and simple interactive elements

You will want something else for:

• High-end brand photography or campaign imagery requiring photorealistic precision
• Complex illustration or original visual identity work
• Animation and motion graphics beyond short generated clips
• High-volume image generation, where the monthly allowance makes a dedicated tool cheaper
• Any workflow that cannot tolerate a feature arriving on a staged rollout

Websites sit in a category of their own. Canva can produce a landing page, but a commercial site needs information architecture, page speed, schema and conversion paths designed around how your buyers actually behave, which is what designing a site properly involves rather than what a template delivers. Pairing Canva for volume content with proper build work for the site itself is the arrangement that works for most SMEs.

Where Canva AI fits in an SME content system

For most SMEs, Canva AI is worth paying for with appropriate expectations. Pro at £100 a year is cost-effective against the alternatives, and having the AI tools inside a full design environment is faster in practice than managing separate AI image and writing tools. Magic Resize alone returns meaningful production time for businesses working across multiple formats.

The calculation changes for teams. At £180 per person per year with no volume discount, Canva Business needs to be justified by the brand governance and admin controls rather than the AI allowance, because the allowance increase alone rarely covers the difference for a small team.

What it does not replace: strategic thinking, brand direction, high-production video, and the creative work that differentiates a brand rather than simply producing content at pace.

ProfileTree’s work with SMEs across Northern Ireland, Ireland and Great Britain consistently shows that the gap between businesses getting value from tools like Canva and businesses not getting value from them is almost never about the tool. It is about workflow design, governance and training. The businesses that get consistent results have set up Brand Kits, trained the relevant people to a shared prompt-writing approach, and fitted Canva into an existing content approval process. The ones treating it as a quick shortcut produce inconsistent visual output that undermines rather than supports their brand.

Used well, Canva AI is a production accelerator. Used without a system behind it, it produces a high volume of content that looks similar to everything else.

Frequently Asked Questions