We built a full platform that matches home sellers with the estate agent that suits them, with a CRM behind it and a brand kit on top, in one month, using Claude as the main builder. In our estimate, that is the equivalent of a £20,000 build. It is our own venture, run with a local personal estate agent, and it is live and handling real enquiries. This is a case study of what we built, what it cost in tools, and what it could mean for anyone with an idea they have been told is too big to start.
The problem: a new idea that looked like months of work
Choosing an estate agent is mostly guesswork. Sellers compare fees and glossy brochures, but the thing that decides how the sale feels is the person: how they communicate, how hard they push, how well they fit you. We wanted a way to match a seller to the agent who suits them, rather than the agent who shouts loudest.
As a brand-new concept, it looked expensive. In our estimate, building a minimum viable product the traditional way would have been the equivalent of a £20,000 build and taken several months. That is our own judgement from experience of agency quotes, not a published figure, so treat it as an illustration of the scale rather than a benchmark. Set against that, the tools for our version cost about £500.
What we built in a month
There are three pieces, and all of them came from the same build:
- A matching tool. A seller answers questions about their sale, such as the type of property and what they need, and about the personality they want in an agent. The tool then matches them to the agent who fits best.
- A CRM for agents and their leads. Every enquiry is tracked in one back end, so an agent can see where each lead came from and what happens next, rather than juggling spreadsheets and inboxes.
- A brand kit. Once the colours, type and tone were settled, we could produce consistent, good-looking marketing material and reuse it across social media without starting from scratch each time.
We have the admin screens too, but they hold lead data, so we are not showing them here. Anything we do share is demo or blurred data.
The tools, and what they cost
The build used Claude to design and write the platform, with a small set of services around it: Postgres for the database, Vercel for hosting, and Apify and Firecrawl for gathering and structuring data from the web. Tool spend came to about £500 to build, and about £90 a month to run.
To be precise about that: those are tool costs only. They are not a fee for anything, and they do not include our own time. The point is how little software spend a working product can now need.
What it could mean for your business
We are not going to turn one project into a promise. But the shape of the result is worth reading carefully:
- Speed to a first version. One month from a blank page to a working product that real people use. That changes what is worth trying, because a bad idea now costs weeks rather than a year.
- A low running cost. About £90 a month in tools means you can keep a niche product alive while it finds its feet.
- Feedback from real users early. The agents involved have given us strong feedback, and the platform has already brought in several leads, most of which went on to instruct. That is a small sample, so we draw no big conclusions, but it is a real signal rather than a guess.
If you are weighing up a tool for your own business, such as a quote calculator, a booking flow, a lead tracker or a matching form, the same pattern could apply. Start with the smallest version that answers your biggest question, put it in front of a few real people, and add to it from what they tell you.
How we would approach it again
Four habits made the month work:
- Describe the outcome first. We wrote down who the tool is for and what a good result looks like before building anything.
- Build the thinnest slice. Matching first, then the CRM, then the brand kit. Each piece was usable before the next began.
- Keep a person in the loop. Agents review and act on every lead. The tool does the matching and the tracking, not the relationship.
- Keep the stack small. Fewer services means fewer things to pay for, learn and fix.
Five questions to ask before you build your own
If this has you thinking about a tool of your own, these are the questions we would put to you in a first conversation:
- Who uses it, and what do they do today? If the answer is a spreadsheet, a stack of emails or a phone call, you have found your first thing to replace.
- What is the one thing it must do on day one? Everything else can wait for version two.
- Who needs to approve what it produces? Decide where a person checks the output before it reaches a customer.
- What will it cost each month? Add up hosting, database and any data services, and compare that with what the problem costs you today.
- How will you know it is working? Pick one simple measure, such as enquiries handled, time saved per week or feedback from the first few users.
None of these needs technical knowledge to answer, which is part of the point. The hard part of a build like this is knowing what you want, and that is a business question before it is a technical one.
What this is, and is not
This is our own project, built for ourselves and a partner. Your costs and results will differ depending on what you build and how complex it is, and a tool like this still needs ongoing care. It does not replace the judgement of the people using it.
If you have an idea you want to bring to life, contact us. A free 30 minute AI audit is a good place to start, and we will tell you plainly what could be built, how long it might take and what the tools would cost.
