
Design Studio for Gutenberg, a WordPress plugin released on 16 November 2026, sells for $39 with per-person unlimited use and a licence that never expires. The price makes it a clean test case for the AI automation for business question that marketing teams now hear in every planning cycle: when production tools keep getting cheaper, which work should a person still do? The plugin's answer is that automation absorbs the repetitive steps while design decisions stay with a designer. That sets the boundary inside the workflow rather than at the studio door.
Key takeaways:
DSG is a system-wide plugin rather than a page builder. The theme stays the theme and Gutenberg stays Gutenberg, while the plugin adds controls across the editor. The plugin's site notes that a Gutenberg page is lighter and loads faster than one built with a page builder, which is the argument for living inside the editor instead of on top of it.
The controls are narrow and specific. A Type Director Block carries kerning and full font and style selection that is not tied to a template, plus keyboard commands of the kind found in desktop design software. An Art Director Block imports multilayer Photoshop or Krita files and treats each layer with blur, invert, halftone, sepia or full blend settings, with its own export settings. A Composition Block adds a height control and lets the other blocks sit anywhere, including overlapping. Rulers, guides, a grid, device-specific visibility and a Designer Dash backend with pre-flight checks fill out the set.
None of those features generates anything. The plugin assumes a person with a point of view is driving, and it spends its complexity on the decisions that person makes.

The number worth taking apart is the unit the price is charged on. DSG is sold once for $39 per person, covers unlimited use by that person, and the licence never expires. The plugin's FAQ puts no cap on production and gives the range as one site to 5,000 sites, so the same $39 covers a single brochure page or a portfolio of client builds. The current release carries the version number Design Studio 1.0.
Set the metric before comparing that with anything else. Cost per published site, over the life of the licence, is the price divided by the number of sites the team builds with it. Because the numerator is fixed and the denominator has no ceiling, the figure falls as the team publishes more. Most AI automation for business comparisons stop at the headline fee and never divide it this way.
The blind spot matters as much as the calculation. This metric says nothing about hosting, maintenance, time to a first publish, or the hours a designer needs to learn the tool. Comparing a licence with a monthly creative subscription on cost per finished asset only works once both sides use the same definition.
This is the point the search results leave open. Pages that explain AI automation for business describe what the technology does, and none of them says where its jurisdiction ends in web and campaign production. The plugin's author draws that line at judgment. In his account, AI should remove drudgery, generate first passes, organise assets and automate anything that gains nothing from a human decision. Craft, he writes, begins where the default answer ends.
The glossary definition of AI automation is a fair boundary marker: large language models and agents execute tasks and multi-step workflows that used to need human judgment, handling ambiguity and context where rule-based automation could not. The example given is a customer email read, its intent inferred, a reply drafted and that draft routed for approval. The drafting is machine work; the approval is not.
Web production follows the same shape. Cropping and resizing a set of images, drafting first-pass page copy, sorting a folder of assets, writing headline variants for a test: each has a cheap wrong answer and a fast correction. When a mistake costs a minute to fix, automating the step is the rational choice.
The news story names the decisions it means. Whether two letters need tighter kerning. Whether an image should move down three pixels. Whether a composition should deliberately break the grid. Whether the mobile version deserves its own layout instead of a shrunken copy of the desktop one. Each of these is a judgement about intent, and each is invisible to a metric that counts pages published.
Written as a question a manager can ask in a meeting, the test is short. Would a competent reviewer spot a wrong answer within a minute? If yes, automate the step and review the output. If a wrong answer would look acceptable to everyone except the person who owns the brand, the decision stays with that person.

| Task | Default owner | What the review checks |
|---|---|---|
| First-pass page copy | Automation | Claims, tone, and whether the draft says anything the brand cannot support |
| Headline setting and kerning | Designer | Optical spacing at display sizes, and line breaks that hold the meaning |
| Image compositing and retouching | Designer, with automated preparation | Layer treatment, colour, and whether the composite survives a crop |
| Mobile layout | Designer | Whether the small screen keeps the hierarchy, not only the content |
| Resizing one design across placements | Automation | Safe margins, platform specs, legibility at the smallest size |
| Campaign variant production | Automation, designer signs off | Consistency across the set, and one variant that carries the idea |
| Seasonal theme or colour switch | Template system | Contrast, accessibility, and whether the palette still reads as the brand |
The table is a starting position. A team with one designer and a heavy campaign calendar will move more rows to automation than a studio with spare capacity, and both can be right as long as the sign-off is named in advance.
Write the metric definition and the time range down before the tool goes in, or the numbers will be read backwards later. Four definitions cover most of it: time from brief to first reviewable draft, review cycles per page, rework items per finished asset, and cost per finished asset. Use a window of the same length before and after the change, and keep the same reviewers inside it.
Correlation is the trap in every AI automation for business review. A shorter cycle after a rollout may come from the tool, from a stakeholder who left, or from a quiet quarter, and a project tracker separates none of those causes by itself. Take one calculation apart instead. If a team publishes twice as many pages in the same window while rework items per page also rise, the automation moved volume and left the craft cost where it was, which is a different result from the one the dashboard suggests.
Privacy belongs in the same conversation. Production data pulled from live sites should be aggregated before it reaches a report, and measuring visitor behaviour on those pages still runs under the consent rules of the market, including the PDPC's guidance in Singapore.
Topkee runs its web and creative work on the same division, and its product notes are explicit about where the line sits. Weber Web Builder lets a team build and publish sites without a designer or a coder: preset templates, drag-and-drop modules, instant preview, brand colours and typography matching, and a one-click theme switch for a seasonal campaign, with built-in SEO tools and one-to-one training so the client keeps editing after handover. The design controls stay in the user's hands instead of being automated away.
Custom AI agent development follows the draft-and-verify pattern. Agents combine knowledge retrieval, workflow orchestration and tool calls with explicit human confirmation points, and anything sent externally or any budget or order change needs approval or pre-authorisation. The delivery principle is stated as sources attributable, execution permissioned, results acceptable, handover possible.
Topkee Creative covers the work that should not be generated: human-made key visuals and videos on a monthly plan, counted per finished design rather than per exported file, across Google, YouTube, Instagram, Facebook, rednote, TikTok, LinkedIn and Yahoo placements. Topkee has no connection to Design Studio for Gutenberg, and the parallel is only about the same boundary drawn in a different business.
A licence suits a team that already holds design judgement in-house and wants better controls. It does not suit a team that needs finished campaign work every month and has nobody to operate those controls. That is the case for buying output rather than software: a video production company working on a monthly subscription delivers finished key visuals and short videos, and the scope is agreed before production instead of being discovered after the first export.
The two purchases answer different questions. One asks what the team can build with better tools; the other asks who builds it. Mixing them up is where budgets go wrong, because a licence bought to solve a capacity problem usually ends with the same queue and a subscription added on top.

Open the live demo and run one real page through it. Set a headline with kerning, composite a layered image, place a block off the grid, then look at the same page on a phone. If the controls feel natural to whoever will use them, the licence is cheap enough to test on a single project. If nobody on the team wants to touch them, the tool will not clear the queue.
Then read two related pieces before you commit the rest of the budget. One covers which parts of a creative tool's AI feature set are worth adopting and which are better left off. The other explains how agent traffic and AI answers change what a page has to earn, which decides how much of the production line can safely be automated.
A team that needs finished campaign work every month should price a video production company retainer against the licence before choosing between them, and a team with design skill in-house should start with the demo.
No. It is a system-wide plugin that adds design controls across Gutenberg while the theme and the editor stay as they are. The plugin's site makes the distinction directly: a page builder was built to make construction easier, while DSG was built to turn the editor into a design tool.
Per-person unlimited use, which means no counting of websites, and a licence that never expires. The plugin's FAQ adds that upgrades and support are available on renewal, so the version already bought keeps working on old and new sites.
The plugin's author asked ChatGPT how its own web-building compared with a competent designer using DSG, and reports the reply that much of what the plugin enables is still awkward to get from AI. That is one vendor's account, not a benchmark. Treat it as a reason to test the claim on your own pages.
Start with steps where a wrong output is obvious within a review and cheap to correct, such as resizing, first-pass copy and asset sorting. Leave the steps where a wrong output still looks acceptable to everyone except the brand owner.
The plugin assumes desktop design habits such as kerning, layers and guides. A marketer can learn the controls, but the decisions those controls exist for are the ones a designer is trained to make. Where nobody holds that judgement, a creative subscription is the more direct purchase.