Figma Make for designers: what changes and what does not

Figma Make turns a text prompt, an image or a Figma frame into a working prototype or app. For designers it speeds up first drafts and compresses simple template work, while brand, product and design system work still depends on human judgment. Price your judgment and treat Make as a fast junior.

What Figma Make does

Figma Make is an AI tool inside Figma that builds prototypes, web apps and frontends from prompts. Figma's Make page says outputs are code-backed and can be edited visually on the canvas. It accepts text prompts, voice dictation and attachments such as designs, images, videos and PDFs, and it can use design context from Make kits (design systems, libraries, guidelines) and Figma frames.

A feature to build Make in a local codebase and ship to production is marked "coming soon" on that page, so check the current state before you promise a client a code handoff.

What that means in practice: a landing page, a pricing page, a simple dashboard or a conventional onboarding flow can reach a decent first draft in a short session. The quality is closer to a templated brief handled by a junior than to senior craft. That is my reading, not Figma's claim, so test it on your own work.

What it costs: credits, not a flat fee

Make runs on AI credits, and the plan you are on sets how many you get. Figma's pricing page lists Professional at $16 per Full seat per month with 3,000 AI credits, Organization at $55 with 3,500, and Enterprise at $90 with 4,250. The free Starter plan has 500 credits a month. Make is included with Full seats.

Credit use is not fixed. Figma's help article on AI credits says use depends on the model, the task and the context you supply, and gives rough examples: changing a font is about 30 credits or more, making a design interactive about 75 or more, and generating an app from scratch about 100 or more. Credits reset monthly and do not roll over.

For a solo designer, that makes Make a variable cost. A client project that needs many generations can use a chunk of a month's allowance, so decide whether to absorb that or build it into your fee.

What the data says about designers and AI

Designers are less convinced than developers. Figma's own 2025 AI report, based on about 2,500 users, found 69% of designers satisfied with AI tools against 82% of developers. It also found that only 31% of designers use AI for core design work such as asset generation, against 59% of developers who use it for core development tasks like code generation.

That gap is useful. It suggests the tools are further along at producing code than at producing design that people trust. It also suggests that many designers still do the central work by hand, which is where your value sits.

On the labour market, the US Bureau of Labor Statistics shows about 253,100 graphic designer jobs with a median annual pay of $62,960 in 2025. I could not find a reliable public figure for how AI has changed design rates, so I will not give you one. Treat any blog that quotes exact percentage changes without a named source with caution.

Which design work looks exposed

Based on what Make does, three kinds of work are the most exposed.

Template-grade pages

Single-purpose landing pages, lead magnet pages, simple blog layouts and basic dashboards. The buyer for these can now get a first draft in minutes, often from an in-house marketer. If this was most of your income, price pressure is a reasonable expectation.

Directions and mockups

"Show me three directions" used to be a paid phase. A client can now generate directions in the meeting. The risk is being hired only after the direction is chosen.

Screen production at scale

Producing many similar screens that vary by plan or empty state is exactly what a prompt plus a design system does well.

Which work looks safer

These are judgment calls, not measurements.

• Brand identity. Logo systems, type, colour and voice need taste and a client conversation. AI drafts tend to look generic.
• Product design at the system level. Flows, states, edge cases and accessibility need someone who understands the product.
• Design system architecture. Make works better with clean, well-named systems, which someone has to build.
• Research, workshops and strategy. They depend on real conversations.
• Illustration and motion with a distinct hand.

If your portfolio leans on the first list, add proof of the second. If it already leans on the second, the tool is more help than threat.

What to do as a designer

Start by using Make yourself. Build a few of your own past briefs with it and see where it fails. You will learn what to charge for.

Then change what you sell. Sell outcomes (more demo bookings, fewer support tickets, a coherent system) and not files. A client paying for "a landing page" now compares you to a prompt. A client paying for "a landing page that doubled bookings, with the testing plan" does not.

Pricing follows. Some work moves from large project fees to smaller editing and polish fees, while specialist work keeps its price. A monthly plan can smooth that out. The post on design-as-a-service pricing shows how to structure one. For the wider shift in client work, see AI agents and repositioning.

Disclose your tools. Tell clients where you use AI in the process, and check any contract clause about AI use before you sign it. If a client forbids AI for compliance reasons, price accordingly.

Finally, polish the experience around the design. When the design itself is harder to tell apart, the way you run the job becomes part of the product. A single branded place for the brief, files, approvals and invoice signals a serious practice, as covered in the best client portal software.

A quick way to test it on your own work

Pick one past project that you billed as a template job. Write the original brief as a prompt, run it through Make, and compare the result with what you delivered. Note three things: how long the first usable draft took, how many credits it used, and how much cleanup it needed before you would show it to a client.

If the draft needed an hour of cleanup, that is your new baseline for fast work, and you can price a "first draft plus polish" tier around it. If it needed a day, the tool is not yet a threat to that kind of job. Repeat the test every few months, because the tool changes quickly and the answer will too.

Keep a short record of each test. It becomes a useful case study, and it gives you something concrete to say when a client asks how you use AI.

How Delivvo supports a design practice that uses AI

Delivvo does not generate designs. It runs the client side, so the time you save with Make does not leak into admin.

1. Collect briefs with an inquiry form at delivvo.io/f/yourname. It has up to 40 fields and a file field, so a client can attach references, screenshots and examples they like, and describe the goal in their own words.
2. The lead lands on your lead board. Reply from there.
3. Send a quote built from packages and add-ons. List "Make-assisted first draft" and "brand system" as separate lines with their own prices. On Pro, the client picks, signs and pays the deposit on one page.
4. Deliver work in the client portal. Each file keeps its versions, so a client who asks to go back to the second direction can. Clients approve from the portal with a click.
5. Bill by recurring invoice for retainers. Late payment reminders go out at 7, 14 and 30 days and stop when the invoice is paid, so you are not chasing money in the hours Make has just given back to you.

Pro allows files of up to 50 GB each, resumable, which suits large design exports and video. Clients open the portal with a one-time code and there is no password. You can add your logo and colour to everything the client sees. The client pays through your own gateway, and Delivvo takes 0%.

Free covers 3 clients and 3 active projects with no card. See the proposals and client portal pages.

Read more about contracts and e-signatures, or start with the CRM and client portal.

FAQ