The Designer Is Still the System

AI is making execution abundant.

Research can become a brief in hours. A component can move from Figma into code without waiting for a traditional handoff. Content, product data and redirects can be moved between platforms at a scale that would make manual work both slow and slightly unkind.

This does not make the designer less important.

It makes judgement more valuable.

Faster does not mean less thinking. It means less time lost between thinking and execution. By using agents for research, production, integration and repetitive technical work, more of the client’s investment goes towards senior judgement, creative development and improving the result.

At Studio Seneca, we use Codex, Claude, Gemini, Figma, GitHub and Framer as one connected workflow. Agents handle research, repetition, code production, data, integrations and migrations. The human remains responsible for deciding what should be built, why it should exist and whether the result is good enough to carry the client’s name.

Productivity is not the number of screens produced in an afternoon.

It is the amount of senior attention returned to the decisions that create value.

Discovery becomes faster, not automatic

We begin in Codex.

Ideas are developed, compared and challenged. Research across the category, customer, existing website and technical environment is assembled more quickly. Gemini provides another research path, while both Gemini and Codex support image generation during visual exploration.

The advantage is breadth. Agents can scan competitors, group customer concerns, identify repeated claims and surface technical questions before the studio commits to a direction.

The limitation is meaning.

Research does not become strategy until someone decides what matters. An agent can identify ten conventions in a category. It cannot decide which convention a brand has earned the right to break.

Human input is present from the beginning. Discovery should challenge the client’s assumptions, not simply organise them into a cleaner document.

Feasibility enters before approval

Design and development have traditionally been separated by a polite but expensive gap.

The design is approved. The development team receives it. Someone then discovers that the elegant interaction requires data the client does not have or an integration the platform does not support.

Codex allows us to test feasibility while the idea is still flexible. We can inspect the platform, content model, APIs, data and likely integration requirements before a design decision becomes costly to reverse.

This does not mean designing only what is easy to build. It means understanding the cost and consequence of the idea before presenting it as inevitable.

Feasibility becomes a design material, not a meeting held after design.

Figma is the decision layer

The work then moves into Figma.

This is where strategy becomes visible and the design DNA is established. Hierarchy, typography, imagery, spacing, interaction and behaviour are considered as one system.

Figma is not the only place where human judgement appears. It is the place where that judgement becomes tangible, reviewable and approved before agents carry it into production.

AI can generate an endless supply of plausible layouts. Plausibility is not a position. A brand does not become distinctive by producing more versions of the same fashionable homepage.

The designer creates coherence. The type has to support the voice. The images have to carry the right emotional temperature. The commercial structure has to work without reducing the brand to a collection of conversion devices.

The human does not arrive at the end to approve what a machine has made.

The human defines the system the machines are asked to extend.

The system lives in Figma and GitHub

Approved components live in Figma. Their coded counterparts live in GitHub.

Figma retains the visual rules through reusable components, variables and styles. GitHub retains their implemented behaviour. Together, they create studio memory.

This reduces the need to rebuild familiar patterns on every project. It also makes more time available for the decisions that should be specific to the client.

The connection still requires review. A button in Figma and a button in code can share a name while behaving differently. AI does not remove drift by itself.

Reuse creates value when the designed and coded systems remain aligned.

Claude is the frontend assistant

In our current workflow, Claude acts as the frontend assistant.

Once the design and component behaviour are approved, it helps translate the Figma system into responsive code. It can build repeated interface patterns, apply components and accelerate the detailed work required to make the design function in a browser.

It does not become the design authority.

Responsive behaviour, accessibility, performance and edge cases still need to be tested. Generated code can appear convincing before it is durable. Studio Seneca remains accountable for what reaches the customer.

The assistant accelerates implementation. It does not own the experience.

Codex is the backend command centre

The visible website is only one part of an ecommerce business.

Behind it sit products, inventory, content, customer data, analytics, email marketing, domains, hosting, payments and fulfilment. These often live in separate platforms with separate interfaces and separate descriptions of the same customer.

Codex operates as the backend command centre.

It helps configure the store, inspect data, connect services and coordinate changes across ecommerce, marketing and hosting systems through one working interface. Each platform remains responsible for its own information, but Codex gives the studio an operational layer across them.

This matters because integration failures are customer experience failures.

A signup form that reaches the wrong list is a design failure. So is a campaign with broken tracking or a product page showing stale inventory. The customer does not care which application created the gap.

The website becomes more than a collection of designed pages. It becomes an operating system for the brand.

SEO and migration belong inside the design process

A new website can look better and perform worse.

This usually happens when migration is treated as an administrative task after the design has finished. Valuable URLs disappear. Redirects are incomplete. Metadata is flattened. Analytics stop measuring the journeys the business depends on.

We use Codex to make migrations more precise, whether the move is WordPress to Shopify, WordPress to WordPress, Shopify to Shopify or WordPress to Framer.

Agents can inventory pages, compare structures, map redirects, identify missing metadata, review structured data and test the new site against the old one. Codex also supports ongoing SEO through content structure, technical checks, metadata, sitemaps, analytics and performance review.

Automation allows us to check more. It does not make risk disappear. Redirects require review, tracking requires testing and search performance requires monitoring after launch.

The objective is not an empty promise of perfection.

It is controlled, testable precision.

One project, fewer handoffs

A typical project can now move through the studio like this:

1. Codex audits the existing WordPress or Shopify site, including its content, search structure and technical environment.

2. Codex and Gemini accelerate research, challenge the initial proposition and explore possible visual territories.

3. Feasibility is tested before the experience is approved.

4. Studio Seneca creates the brand direction, interface and component system in Figma.

5. Human review establishes what is approved and what must change.

6. Claude supports the responsive frontend implementation.

7. Reusable code is retained in GitHub.

8. Codex connects the store, marketing, analytics and hosting systems.

9. Migration checks protect URLs, metadata, content, product data and tracking.

10. The studio reviews, tests and takes responsibility for launch.

The tools are not interesting because there are several of them. They are useful because the decisions survive between them.

Ecommerce and branded experiences take different routes

Most of this workflow is designed around ecommerce.

Commerce requires operational depth. Product data, inventory, payments, promotions, marketing and analytics have to work behind the branded experience. A beautiful interface cannot negotiate with a broken checkout.

For more expressive, brand-led websites, we continue to use Framer. It provides a direct route from visual design to a responsive, published experience and is especially useful when storytelling, movement and the feel of the page carry more of the value.

This is not a contest between platforms. It is a decision about the problem.

AI makes both production paths faster. It does not make them interchangeable.

The studio changes with the workflow

AI allows a smaller senior team to carry an idea further.

Strategy can include technical reality. Design can anticipate the data behind the interface. Frontend implementation can remain closer to the approved system. Backend operations, marketing integration, SEO and migration can be considered as parts of the same customer experience.

This changes the commercial model of the studio.

The value is no longer the number of production hours required to move information between applications or recreate familiar components. It is the quality of the decisions, the range of capability connected to them and the standard of the system delivered to the client.

The studio should not sell the hours automation removes.

It should sell what becomes possible because those hours have been returned.

AI can accelerate mediocrity too

Speed does not naturally produce quality.

AI can generate many competent but interchangeable options. It can repeat a weak assumption across an entire website, turn a fashionable visual device into a design system and introduce migration mistakes at a scale that would be difficult to achieve manually.

The danger is not only bad output.

It is polished output that feels finished before anyone has asked whether it is right.

This makes the designer more important, not less. Taste, context and empathy remain scarce. Accountability is scarcer still.

An agent can recommend, generate and execute. It cannot take responsibility for the client’s position, customer experience, search performance or commercial result.

The studio decides what goes live and owns the consequences.

The lesson

AI is collapsing the distance between an idea and its execution.

In our current workflow, Codex expands discovery, tests feasibility and acts as the backend command centre for integrations, SEO and migration. Gemini supports research and visual exploration. Figma is the decision layer and home of the design system. Claude assists with frontend implementation. GitHub retains reusable code. Framer provides a direct production path for branded experiences.

The tools will change. The principle will not.

The future studio will not be the one that produces the most screens in the least time. It will be the one that makes better decisions and carries those decisions further.

The agents extend the system.

The designer decides what the system is for.

What this means for our clients

Our clients do not benefit because a repetitive task took the studio less time.

They benefit when that time returns to the outcome.

They see technical constraints earlier, working versions sooner and more opportunities to improve the experience before launch. Strategy, design and development remain connected, so fewer decisions are diluted through handoffs. SEO and accumulated search value receive greater protection during migration. The store, marketing, analytics and hosting systems are considered as parts of one customer journey rather than a collection of unrelated applications.

The result is not a website produced more cheaply by machines.

It is a better considered brand experience, supported by a more capable operating system and built to keep evolving after launch.

AI changes the studio behind the work. For our clients, that means less time lost between an idea and its execution, and more attention given to the decision that makes both worthwhile.