Twelve browser tabs for one address
A new enquiry for a house in a town you have not worked in before. Before the feasibility study, someone needs to know the site's context. Is it in a conservation area? Is the house or a neighbour listed? Is there an Article 4 direction? Any tree preservation orders? What flood zone is it in? What applications have been made on this property and its neighbours, and what happened to them?
So a part 2 architect opens the council's planning search, then the council's map of conservation areas, then the national listed building register, then the flood map, then the council's tree preservation order page. Each council's website is different. Some have maps, some have PDFs. An afternoon later, there is a page of notes, and the architect is still not sure they have found everything.
Why this is still manual
The information is public, but it is fragmented. Each council runs its own planning register and constraint maps, often on different systems. National datasets cover some designations but not others, and not always up to date. There is no single place to type an address and see everything, so each practice builds the picture by hand for every site.
- Planning registers differ by council and are searched one at a time.
- Constraint maps are in different formats, some only as PDFs.
- Neighbouring applications, often the most useful precedent, are hard to find.
- Findings are written up in notes that vary by person.
What slow site research costs
Feasibility time is often unpaid or on a small fee, so hours spent on research come straight off the margin. A missed constraint, such as an Article 4 direction removing permitted development rights, can mean advice to the client that has to be corrected later. And the same council gets researched again from scratch on the next job in that area.
A site pack from an address
- The architect enters an address or drops a pin, and the site boundary is identified from public mapping data where available.
- Designations are checked against national and council datasets that are published in usable form, such as conservation areas, listed buildings, flood zones and, where published, tree preservation orders and Article 4 directions.
- The council's planning register is searched for applications on the site and nearby, and results are listed with the decision, date and a link to the documents.
- A language model can summarise the decision notices of relevant nearby applications, such as what was approved and any common reasons for refusal, with links to the originals.
- The pack is produced as a single document or page, clearly showing which sources were checked, when, and which could not be checked automatically.
- The architect verifies the findings, adds their own notes and files the pack to the project.
- Each council's quirks are captured once, so the next site in the same area is quicker.
The pack is a research aid. Designations and planning history should be confirmed with the local planning authority, and planning judgement stays with the architect and, where needed, a planning consultant.
What the pack covers
| Item | Source type | Checked automatically? |
|---|---|---|
| Conservation area | National or council data | Where published |
| Listed buildings nearby | National register | Yes |
| Flood zone | Government flood map data | Yes |
| Article 4 and TPOs | Council data | Where published |
| Planning history on and near the site | Council register | Where searchable |
Feasibility that starts from the full picture
The architect begins the feasibility with the constraints and precedents in front of them, and with a clear list of what still needs checking by hand. Advice to the client is grounded from the start. And the pack becomes part of the project record, useful again when the planning application is prepared.
Is this your feasibility process?
- Site research means an afternoon on council websites.
- Each architect records findings in their own way.
- Nearby planning applications are hard to find.
- A constraint has been missed at feasibility before.