What 300 Hours Building an AI Storefront Tool Taught Me About Product-Market Fit

Over the past few months, I have spent around 300 hours building an AI storefront visualisation app called Streetscape.
It has been one of those projects that started with a clear idea, grew into a working product, attracted interest from serious property companies, and then taught me something I probably needed to learn the hard way.
The product worked.
That is the strange part.
The images were strong. The generation flow worked. The app had login, signup, account handling, trial access, PAYG thinking, subscription tiers, Firebase, Vercel deployment, and the small but important details that turn an idea into something real.
There was genuine interest too. British Land and Landsec both took the time to look at it and trial the software, which I appreciated. When companies of that scale are willing to engage with a new idea, it gives you confidence that you are working on something relevant.
But interest is not the same as uptake.
That has probably been the biggest lesson.

The original idea

Streetscape was built around a simple problem.
Retail property teams often need to visualise how a unit, frontage, hoarding, leasing concept or placemaking idea might look before committing to design, CGI or agency production.
Traditionally, that can be slow and expensive. If you need a full CGI view, the cost can quickly run into the thousands. In one recent project for another client, a visual route moved into CGI because the exact base view did not exist. That work came in at around £1,800.
That is a valid route when you need a view created from scratch.
But if you already have a photograph or reference image, AI can be a much faster way to explore the visual direction. It can help answer early questions:

Could this unit look better with a different frontage?

Would this concept work on a real street?

How might a hoarding design feel in context?

Is this idea worth developing further before spending properly on production?

That was the opportunity I saw.

What I built

The first version of Streetscape was not just a set of prompts or mockups. It became a proper web app.
It included user accounts, login and signup flows, image generation, trial access, pricing logic, subscription tiers, PAYG thinking, backend services, deployment, and a lot of testing.
There were plenty of technical challenges along the way. Some were expected. Others appeared only when real users started testing. That is the nature of software: the moment it becomes real, the edge cases arrive.
The app became stronger through that process. Login messaging improved. Signup and onboarding were hardened. Firestore provisioning was moved to a more secure server-side flow. The product became more robust.
From a build point of view, it was a valuable project.
From a commercial point of view, it was more complicated.

The subscription model did not get enough pull

The original commercial idea included subscription tiers for teams and enterprise users.
On paper, that made sense. If a property company had recurring visualisation needs, a subscription could provide regular access and predictable usage.
But the market did not respond strongly enough to that model.
There was interest. There were conversations. There were trials. But there was not enough uptake to justify continuing to develop the subscription side of the product.
That is a difficult thing to admit after putting in so much work.
It is tempting to think the answer is another feature, another pricing page, another trial option, another email, another product tweak.
Sometimes that is true.
But sometimes the more honest answer is that the buying behaviour is not what you thought it would be.

PAYG feels like the better fit

The more I looked at it, the more I felt that Streetscape fits better as a PAYG tool than as another monthly subscription.
That does not mean the product has no value. It means the value may be tied to specific moments rather than ongoing platform access.
A property team may not need to generate storefront concepts every week. But when they do have a live unit, a leasing pack, a hoarding idea, or a placemaking proposal, a quick AI visual could be useful.
That is a different kind of buying behaviour.
It is more project-led. More occasional. More practical.
So I have decided to pause the subscription tiers and keep Streetscape available on a PAYG basis.
That feels more realistic.

The lesson

The biggest lesson from all of this is that building the product is only part of the job.
Finding the right buying behaviour is the other part.
You can build something useful and still have the wrong commercial model. You can get likes, interest and positive feedback without creating enough urgency for people to pay. You can solve a real problem but still need to package it in the way customers actually want to buy.
That is not failure. It is market feedback.
But it is hard feedback when you have invested a lot of time and energy.
For me, the rule now is simple:

No more speculative development.
Streetscape will stay alive, but future work needs to be tied to real use cases, real projects, or real demand.
That means PAYG access, client-led visuals, and practical applications where the value is clear.

Where Streetscape sits now

Streetscape is not closing.
It is changing shape.
The subscription route is being paused. PAYG remains open. The product will continue to be available for situations where a client has a real photo or reference view and wants to explore a storefront or retail concept quickly.
If a base view does not exist, CGI still has its place. But if the view already exists, AI can be a faster and more cost-effective way to test the visual direction before committing to heavier production.
That is where I think the real value sits.

Sometimes the win is not the product you thought you were building.
Sometimes the win is learning where the value actually is.

Work with us

If you work in retail property, leasing, regeneration, placemaking or high street development and have a live unit or concept you would like to visualise, Streetscape is available on a PAYG basis at https://www.streetscape.ws 
Bring a photo or reference view, and we can test the idea quickly before you commit to a larger design or CGI route.