The hidden work behind BCAF’s 2027 expansion

BCAF’s brief for 2027 sounded simple: add two new types of exhibition space so more artists could take part. But when a platform manages applications, artist records, inventory, payments, and admin dashboards simultaneously, a simple feature request touches almost every piece of underlying architecture.

When a successful system reaches its limit

The 2025 and 2026 fairs had both sold out, with room for somewhere between 40 and 50 artists depending on the layout. For 2027, BCAF wanted to grow, adding two new walkway formats and taking total capacity from 40 spaces to 100.

That is a good problem for an event to have, but it is exactly the kind of change that exposes the limits of a system built for a smaller scale. A website can work perfectly well for years and still start to strain the moment the organisation behind it outgrows what it was originally designed to handle.

Why two new space types touched the whole platform

“Add two more space options” is a simple request, but it isn’t necessarily a simple change. The new space options had to behave consistently everywhere they appeared:

• as choices in the artist’s booking journey
• as entries that fed correctly into availability and stock
• as products tied to the right event dates
• as items that carried through cart, checkout and payment without dropping any detail along the way.

They also had to show up properly in artist dashboards and event status reporting, respect the rules around discounts and multi-event bookings, and be visible and controllable from the admin side, including for manual bookings.

None of that is optional. Miss one connection and a booking can look complete on screen while quietly breaking somewhere between availability, payment and the artist’s record.

From one-off fixes to a reusable structure

Rather than bolt the new spaces on as a one-off change, we used the update to build a more structured way of setting up each year’s art fair, connecting event dates, products, space types and availability more consistently so that preparing future years means working through clear processes rather than repeating disconnected changes across different parts of the site.

That is the kind of groundwork that doesn’t show up on the surface, but it’s what matters most the next time the event grows.

Strengthening how applications are handled

The update also improved the operational side that BCAF’s admins deals with directly. Applications now move through clearer states, covering new, approved, rejected, duplicate and suspended cases, each reflecting how the process actually works rather than treating every returning email address the same way.

A rejected artist can apply again, an approved one is guided towards registration rather than submitting a fresh application, and a duplicate or suspended case gets its own, more careful handling.

Manual booking actions were tightened too, restricted to administrators and backed by stronger validation, so there’s less room for a change to end up in the wrong place.

How AI-assisted development changed the workflow

AI-assisted development played its part in getting through this quickly. It helped us review the existing code, trace how events, products, spaces and bookings connected to one another, spot repetitive patterns worth consolidating, and produce first-pass implementations for JavaScript interactions and edge cases that we could then test and refine ourselves.

What it didn’t do was decide how BCAF’s application rules should work, or judge which actions belonged behind an administrator login. Those calls came from understanding the client’s business, and every AI-assisted suggestion still went through our own review, testing and debugging before it reached production.

The update took roughly a week of focused development. Given the number of connected workflows involved, we’d normally expect work of this kind to take three to four times as long. That’s an internal estimate rather than a controlled benchmark, but it reflects the practical difference AI-assisted development can make when it’s paired with experienced technical oversight rather than used in place of it.

Testing across the whole journey

Speed only counts if nothing breaks, so we tested across the full booking journey:

• selecting a new walkway space
• checking availability
• adding it to the basket
• working through checkout and payment
• applying discounts
• updating inventory
• confirming bookings on the artist dashboard
• event status reporting
• manual bookings
• application handling

These all reflected the change correctly, on mobile as well as desktop. A new product can look right on its own page and still fail to update availability, or reach the basket without carrying the right event information into checkout, and that’s precisely where this kind of testing earns its keep.

The outcome

BCAF now has a stronger structure behind its 2027 expansion. Capacity has grown from 40 spaces to 100 across two new walkway formats, all built on the same booking, availability and payment logic as the existing options.

Application administration is clearer, manual bookings are better controlled, and preparing future fairs should mean working through a more repeatable process rather than starting again from scratch.

The value here isn’t speed for its own sake. It’s that AI-assisted development, combined with technical judgement, testing and a clear understanding of how BCAF operates, turned what looked like a small feature request into a platform that’s genuinely easier to grow with.

If your own site is starting to show the same strain as your event, membership or user base grows, we’re always happy to talk through what a similar update might look like for you.