Case study · Prototype

Wren: forms your organization controls

A form builder Banjo5 designed and prototyped with AI-assisted development, and planned from day one around data ownership, security and governance.

Why build another form tool?

We've used and tested around 11 form tools, and every one of them was missing something. Some have a weak form builder. Some make integrations hard to set up. Some offer little control over themes and branding. Others make building the form itself a frustrating experience.

Whichever tool we picked, there was always a headache somewhere, and we ended up working around the tool instead of with it.

Wren brings all those lessons together: one form builder designed to check every box, with the organization in control of its data, its branding and its integrations.

What Wren does today

A working builder and preview, with the product's direction already visible.

Visual form builder

Sections with up to four columns, and text, long-text, number, dropdown, radio and checkbox fields.

Conditional logic

Show or hide fields, text and whole sections based on earlier answers, with AND/OR rule groups.

Rich content

Formatted text blocks, links and images inside forms — cleaned so they can't run hidden code.

Brand themes

Colors, fonts (including your own), logo, layout width, header and button styles.

Responses & analytics

Response tables, charts and cross-tabulation, explored in the prototype with sample data.

Integration groundwork

Stable field keys for webhooks and APIs, and form version history, demonstrated in the browser.

Building a form with sections and columns.
Showing a section only when it's relevant.
The respondent's view, in the organization's own branding.

An honest prototype

We'd rather show clearly what's built than overpromise.

Working today

  • Designing, theming and previewing complete forms
  • Conditional logic across fields and sections
  • Analytics and response views, using sample data
  • Everything stays in your own browser. Nothing is sent to a server.

Next on the roadmap

  • Publish forms and collect responses, protected against spam and bots from day one
  • Webhook-only mode: answers go straight to your own systems, and Wren keeps no copy once they're delivered
  • Salesforce and Zapier integrations that take minutes to set up
  • Choose where your data is stored, such as the EU or US, to meet GDPR and similar rules

AI-assisted, human-governed

AI wrote most of the code. We set the direction, the guardrails and the standard it had to meet. It's the same approach we bring to clients' AI projects.
Step 01

Plan governance first

Before any customer data exists, the roadmap sets out data separation enforced by the database, bot protection, GDPR and data residency.

Step 02

Build fast with AI

With the guardrails set, AI coding tools turned ideas into a working builder in a fraction of the usual time, so we could test real workflows early.

Step 03

Review what AI wrote

A security review found and fixed unsafe handling of formatted text, and added phishing and webhook protections to the plan.

Step 04

Launch only when it's safe

Sign-ups stay closed until the backend, abuse controls and a full security review are complete.

Lessons for any AI-built product

Speed isn't the risk; skipping review is

AI-written code can look finished long before it's safe. A focused review caught a problem that would only have shown up after launch.

Write the rules down early

A written roadmap covering data separation, privacy and abuse gives AI tools, and people, a standard to build to.

Be clear about what's real

Labeling demo features as demos keeps trust intact, for users, clients and your own team.

Building something with AI?

We help organizations move fast with AI without losing control of their data, their security or their standards.

Let's Talk
Services AI Solutions How We Engage Who We Serve About Let's Talk