PropTech, AI
//
2025
Client
ScanSan Properties
Role
Founding Designer
Timeframe
2022 onwards
Tools
Figma

THE PROBLEM
Answering one question about one postcode meant chasing public sources for days, and sometimes for weeks.
Agents preparing a valuation, investors screening areas, builders checking a plot and first time buyers all asked versions of the same questions. The data existed, spread across government, statistics and mapping sources, each with its own format and its own gaps. PAA had to turn that into one answer a professional could put in front of a client.
ScanSan sells to professionals, but most people meet PAA on the free tier, often with a single postcode in mind. The product had to serve both without making either feel like the wrong customer.
Agents believe in AI, but they haven’t felt it yet
Share of UK estate agents who agree with each statement
Design implication: the first answer has to prove the value on its own. Belief is already there; experience isn’t.
Source: GetAgent survey of 622 UK estate agents, reported by The Negotiator, March 2026
Adoption is wide, but shallow
Most firms are experimenting with AI; few use it every day
UK letting agents
Goodlord 2025 · n = 350+
Commercial property firms (global)
RICS 2026 · n = 1,265
Design implication: PAA has to earn a place in a weekly routine (saved areas, monthly reports), not just a good first demo. RICS respondents name data quality and availability as the top barrier, so every answer shows its source and date.
Sources: Goodlord, State of the Lettings Industry 2025 (Oct 2025), 350+ UK letting agents · RICS, AI in Commercial Property and Construction Report 2026 (Q1 2026), 1,265 commercial property respondents, global
WHAT I FOUND
Agents, investors and buyers rarely want a dashboard. They arrive with one question, the rent in a district or the crime near an address, and leave once they have it.
The public data behind those answers is uneven as well. There is no single national register for HMO licences or planning applications, so every section has to say what it knows, what it does not, and how old it is.


THE DECISION
Analytics became one page with eight tabs. Overview, Sales, Rent, Demographics and Crime are free, because that is what an occasional user asks. Growth and trends, Investment and Compare are Pro, because that is what an investor bills for.
Property Check follows the same split, and Pro costs £15 per user a month, or £12 billed yearly. Inside Pro, charts open on the last six years, because property moves in months and years; the full 30 years of data sits one click away.


THE SOLUTION
Tabs, answer actions and paywalls behave the same way wherever they appear, on desktop and on the phone.
Share, PDF, Add to Categories and Bookmark sit at the foot of each answer rather than in the top bar, so a conversation about three districts gives three things to save. The monthly report toggle sits next to the title it belongs to, and every figure carries its source and date.
The Overview tab summarises the other seven with one number and one line each, and every tile opens the tab behind it. The Pro tiles ask the question their tab answers: where is it heading, is it worth buying, which area wins. Property Check reuses the same pattern, with section tabs across the top and Pro rows that unlock in place.

THE BUILD
Seventy-six desktop high fidelity screens across nine stages, built from the ScanSan design system tokens and audited against WCAG 2.2 AA before handoff. B2B comes first, but agents work between viewings, so PAA also ships as a progressive web app: navigation in a side drawer behind the burger menu, the chat pinned to the bottom, and tabs and cards reflowed for one thumb.

THE OUTCOME
Within six months of release, 230 estate agents were using PAA, and 23 property investors had been onboarded and were scanning areas with it to drive their decisions. In the user survey, 95% of users said they were satisfied, and most described the information as highly accurate.
The number that matters most is the time. Building a report or chasing the facts on a postcode went from weeks to minutes or hours. A rating says people liked the product; time given back says it changed how they work.
Agent and investor numbers from ScanSan account data, first six months after release. Satisfaction and time saved from a post release user survey.
Previous case study
Next case study