It's a few years in the future. We're invited to submit our architectural credentials and fees for a project. It's an intriguing project and complex enough to encourage us to search for innovative solutions which could reward our client financially and reputationally. That can also be good for us- we can get some good publicity while at the same time letting our team stretch our problem solving and design skills instead of just relying on our AI tools.

The planning authority has a 'conforming projects' stream in which their AI assessment tools quickly assess applications. To make that smoother the design possibilities in the planning scheme are expressed as metrics. Because outcomes are predictable there is no need for third party appeal. The costs and time of the conforming projects stream planning approvals are therefore greatly reduced.

Our AI design tools already incorporate the codified planning requirements and can mimic the assessment logic of every planning authority's AI assessment tool. We can therefore very quickly create a design which will be immediately approved.

That design is unlikely to be the optimal outcome. But it will be good enough.

If we wish to pursue an innovative design solution which may question those codified planning requirements, we can submit that scheme for approval in the planning authority's 'non-conforming projects stream'. In this stream our scheme is assessed by a human design review panel.

The innovative approach in which we seek to optimise the possible design is inextricably associated with time. To get to such an optimal design with its consequent rewards we will need our team to spend significant time bringing their diverse experience and conceptual skills to iteratively connect new ideas and work across contexts. The subsequent exploring and discussing of the proposal with the design review panel will take even more time.

It is a private sector project, and our client will be balancing possible future value against known hard costs. Finance charges and holding costs have already started and will grow significantly if it takes too long to get planning approval. Because of the time it takes, and the number of humans involved, the non-conforming stream is going to cost our client a lot more than using our AI-based conforming projects approach.

We are thinking about this as we are putting together our submission to our potential client. We look at the huge difference in time and resultant costs between an AI-delivered 'good enough' approach and an optimised and high rewards human-derived scheme.

Should we be surprised if our potential client concludes that 'good enough' is simply good enough? Probably not. But should we as a society be satisfied with 'good enough'?