When people need to click through
Product prototypes
A clickable flow, a small application, or a new feature added to something you already use. Useful when stakeholders need to see how something will actually work.
AI prototyping
AI prototyping to test ideas before a full build
When a new feature, workflow or page needs testing, a slide deck is not enough. We build working prototypes you can click through, share with colleagues, and use to decide whether the idea is worth taking further.
Tools
The tools we use
We use tools such as Lovable, Replit, Claude Code, Codex and Cursor depending on what will produce a working prototype fastest.
Lovable
Replit
Claude Code
Codex
Cursor
Who this is for
Teams with an idea that needs proving
Agencies and marketing teams who want a rapid web prototype or MVP without committing to a full production build straight away.
- Stakeholders need to see something working, not a presentation.
- You want to test whether AI or automation will actually save time.
- You need a prototype before scoping a full website design and build.
What we can build
Working examples, not presentations
When automation might help
Workflow prototypes
Drafting, summarising, sorting or routing work where AI can help but a person still makes the final decision. Useful when you want to test whether automation will actually save time.
When the team needs a tool
Internal tools
Research helpers, first-pass content tools, or quality checks connected to systems such as Notion, Slack or your CRM when that is part of the job.
When the idea is worth keeping
Handover
When a prototype proves useful enough for production, we document how it works, handle the basics properly, and leave you with something your team can understand and maintain.
Have a repetitive task or idea you want to test properly?
How it runs
From idea to clickable prototype
Scope
We agree what the prototype needs to prove and what finished looks like for this stage.
Build
We use the fastest appropriate tools to produce something you can click through and share.
Review
You test with stakeholders and decide whether to iterate, productionise or stop.
Next step
If the idea is worth a full build, we can hand over to website design and build with clear documentation.
Use AI here
Where it tends to help
- Drafting and summarising, followed by a human review.
- Sorting, tagging and routing when the rules are clear.
- Boilerplate code and one-off scripts, with proper checking.
Skip it here
Where it usually does not
- Final client copy that must sound exactly like your brand.
- Decisions that need deep team or client context.
- Tasks a simple checklist or rule already handles well.
Questions
Common questions
- What tools do you use for AI prototyping?
- Lovable, Replit, Claude Code, Codex and Cursor depending on what will produce a working prototype fastest for the job.
- How long does a prototype take?
- Many prototypes are a matter of days once scope is clear. We agree what success looks like before build begins.
- Can a prototype become a production site?
- Sometimes, but often it is better to rebuild properly once the idea is proven. We advise honestly and can take the work into website design and build if needed.
- Do you prototype for agencies as well as direct clients?
- Yes. We often join agency teams to test an idea for a pitch or client presentation before full delivery is scoped.
Related
You might also need
Have a use case in mind?
Tell us what is repetitive, slow or unclear. We will advise whether AI is the right approach, or whether a simpler solution would do.