CRM Providers
How the platform cut property setup time from 15 minutes to 90 seconds
1,400 estate agents manually typing property details into CRM. A 23% error rate. Nine API calls changed the product entirely.
The challenge
The crm platform was built for estate agents who needed a clean, professional way to manage instructions from instruction to completion. Over four years, they'd grown to 1,400 users across 180 branches. The product worked. But one workflow had never been properly solved: adding a new property instruction.
When an agent took on a new instruction, they'd open a new property record in the platform and begin typing. Address. Bedrooms. Property type. Floor area. Tenure. Council tax band. EPC rating and score. Each field had to be populated manually, from memory or from the vendor's paperwork. For experienced agents working quickly, the setup took 12 minutes. For newer team members still finding their way around the system, 15 wasn't unusual.
The error rate was the deeper problem. A 23% data error rate — discovered through an internal audit that cross-referenced CRM entries against Land Registry records — was causing downstream failures across the business. Property portal exports with incorrect floor areas. Compliance reports with wrong EPC ratings. Marketing emails with the wrong bedroom count. Every error that escaped into production created remediation work somewhere else in the business.
Priya Nair, CPO, had been tracking NPS for three years. The manual entry workflow was consistently cited in detractor feedback. The fix was obvious — auto-fill from a reliable data source. The question was finding an API that covered all the fields at a cost that made sense at scale for a CRM that charged per-user SaaS pricing.
"Property intelligence auto-fill went from a nice-to-have to our most-talked-about feature in three months. It didn't just improve data quality — it changed how agents feel about using our product."
The solution
Nine API calls on address selection. Sub-400ms. Bedrooms, property type, floor area, tenure, council tax band, EPC rating and score — all auto-filled before the agent has moved to the next field.
the platform built the auto-fill feature as a four-step pipeline triggered by address selection in the instruction creation form. As soon as an agent selects an address from the autocomplete dropdown, the pipeline fires in the background — the agent sees nothing but a brief loading state before the form populates.
Step one is address resolution: the address find endpoint (/api/public/address/find/, 2 calls) and address retrieve (/api/public/address/retrieve/, 5 calls) confirm the address and return a UPRN. Step two fires two calls in parallel: the property record endpoint and the EPC endpoint.
The property record (/api/v1/properties/{'{uprn}'}/) returns bedrooms, property type, floor area, tenure, and council tax band. The EPC endpoint (/api/epc-checker/{'{uprn}'}/) returns the current energy rating (A–G) and efficiency score. Together, they populate every key field in the instruction form. Nine API calls. Under 400ms. The entire form fills before the agent has finished reading it.
The impact was immediate and measurable. Data errors dropped from 23% to 2.1%. Setup time fell from 12–15 minutes to under 90 seconds. NPS improved by 26 points — the highest score in Their four-year history. In the three months after launch, auto-fill was cited as the number one reason for new sign-ups in post-sale surveys. Among the 1,400 users who used the feature, churn in the 6 months post-launch was zero.
/api/public/address/find/
Autocomplete from typed address. Returns suggestions with UPRN. 36M UK addresses.
/api/public/address/retrieve/
Full Royal Mail formatted address and confirmed UPRN for the selected result.
/api/v1/properties/{'{uprn}'}/
Bedrooms, property type, floor area, tenure, council tax band, lat/lng. Auto-fills 5 form fields.
/api/epc-checker/{'{uprn}'}/
Current energy rating A–G and efficiency score. Auto-fills EPC fields. Always current — never stale.
Address find (2) + retrieve (5) + property record (1) + EPC (1). All completing before the agent has moved to the next field.
The results
Integration
The four-step pipeline runs server-side: address find (2 calls) triggers on each keystroke in the address search field with debouncing; address retrieve (5 calls) fires on selection confirmation; property record (1 call) and EPC (1 call) fire in parallel immediately after UPRN resolution. All calls are proxied through Their backend to protect the API key. The entire pipeline completes in under 400ms on UK infrastructure. At 1,400 active users averaging 4 new instructions per week, the platform runs approximately 22,000–25,000 calls per month — well within the Growth tier. The feature took Their engineering team 6 weeks to build, test, and ship to production.
"Property intelligence auto-fill went from a nice-to-have to our most-talked-about feature in three months. It didn't just improve data quality — it changed how agents feel about using our product."
Read more customer stories
See how estate agents, proptech platforms, surveyors, and mortgage brokers use Homedata.
All customer stories →Get your free API key
100 calls/month free. No credit card. Build your own property auto-fill in a day — the full pipeline is 9 calls.
Start building freeFree tier · No credit card required