Case study · Property
Call the likeliest buyers first
Property developers receive many more enquiries than their sales team can call quickly. This is how we scored leads from enquiry details and engagement in the CRM, so agents prioritise the buyers most likely to reserve, with the model's reasoning visible.
Project overview
Lead prioritisation inside the CRM.
- Industry
- Property
- Client
- Confidential
- Solution
- Lead scoring
- Platform
- Azure ML + CRM
- Timeline
- 6–10 weeks
- Team
- 3 people
- Signals
- Source, budget, engagement, timing
- Output
- Score and reasons in the CRM
- Measurement
- Conversion by score band
Expected business outcomes
What the client set out to achieve
Measured as conversion by score band on real leads.
-
2–4×
Higher conversion in the top score band
Top-scored leads convert several times more often than average.
Target; validated on historical and live leads
-
Faster
Response to high-intent leads
Agents call the top band first.
Where call lists are sorted by score
-
Visible
Why each lead scores high
Main factors shown next to the score.
Model explanations in the CRM
Business challenges
Where the client started
Every enquiry looks the same in the inbox.
Volume
More enquiries than agents can call the same day.
Order
Leads worked in the order they arrived.
Quality
Many enquiries are browsers; some are ready to buy.
Data
Useful signals in the CRM, never used together.
Result
Ready buyers wait while agents call browsers.
Proposed solution
What the build covers
A simple, explainable model inside the tools agents use.
-
Training data
Past enquiries and outcomes prepared from the CRM.
-
Scoring model
Probability of reservation from lead attributes and behaviour.
-
Explanations
Top factors shown next to each score.
-
CRM integration
Scores and bands on lead records and call lists.
-
Retraining
Model retrained monthly on new outcomes.
-
Measurement
Conversion by score band tracked over time.
Project requirements
What the client asked us for
The request was to focus agents' time. The model must be explainable, or agents won't trust it.
We avoid personal characteristics that could lead to unfair treatment, and use behaviour and enquiry details instead.
What the project had to deliver
- 01Scores in the CRM
- 02Reasons agents understand
- 03Fair, behaviour-based signals
- 04Measured conversion
Key challenges & considerations
What made it harder, and what we did
-
Outcome data not recorded in the CRM
What we did
Outcome recording was made a required step in the CRM before the model was trained.
-
Few conversions for small developments
What we did
Developments were pooled where a single one had too few sales to learn from.
-
Market changes affecting behaviour
What we did
The model was retrained on a schedule and monitored for drift.
-
Fairness and privacy requirements
What we did
Unfair signals were excluded, and every score came with its reasons.
What we watched for
- Record outcomes in the CRM
- Explain every score
- Exclude unfair signals
- Retrain as the market changes
Development process
History, model, CRM, measure
Validated on past leads before agents see scores.
-
Data
Enquiries and outcomes extracted and checked.
-
Model
Model trained, validated and checked for fairness.
-
CRM
Scores and explanations added to lead views.
-
Pilot
One sales team uses scores for a month.
-
Rollout
All teams with monthly retraining.
Technology stack
ML inside the CRM
Scores delivered where agents already work.
Machine learning
- Python
- scikit-learn / LightGBM
- Azure Machine Learning
Data
- Microsoft Fabric or Azure SQL
- Feature pipelines
Serving
- Batch scoring
- REST API
- ERP/CRM integration
Monitoring
- Accuracy tracking
- Drift alerts
- Retraining schedule
Quality assurance & deployment
Measured against a test set before launch
AI quality was measured, not judged from a demo. A test set of real examples with agreed answers decided whether the system was ready, and the same set was re-run after every change.
- A test set of real examples with agreed answers
- Accuracy thresholds agreed with the business before launch
- Safety checks: refusals, permissions and prompt-injection tests
- A person reviewing low-confidence output
- Cost and response time measured at real volumes
Deployment
- 01PilotOne sales team uses scores for a month.
- 02RolloutAll teams with monthly retraining.
- 03Pilot, then staged rolloutA small group of real users first, with results reviewed weekly before widening access.
- 04Monitoring and warrantyAccuracy, drift, usage and cost tracked after launch, and six months of free bug fixing.
Final deliverables
Agents' time on the right leads
Agents start each day with a call list ordered by likelihood to buy, with the reasons visible, and ready buyers are contacted first.
Conversion by score band shows the model keeps earning its place.
- Scored leads
- Explained scores
- Prioritised call lists
- Tracked conversion
Handed over to the client
- 01Production AI serviceDeployed in the client's own cloud tenant
- 02Source code and modelsIn a repository the client can access
- 03Test set and resultsThe evidence behind the launch decision
- 04MonitoringAccuracy, usage and cost
- 05Six-month warrantyFree bug fixing after delivery
Success metrics
How success is measured
| Measure | Outcome | How it is measured |
|---|---|---|
| Higher conversion in the top score band | 2–4× | Target; validated on historical and live leads |
| Response to high-intent leads | Faster | Where call lists are sorted by score |
| Why each lead scores high | Visible | Model explanations in the CRM |
Verified reviews
What clients say about our custom software work
Verified reviews from our clients on the same kind of work, published on Clutch.
-
Custom Software Development
Web-Based Coupon App Development
The work produced by the team was of exceptional quality.
Jon Thies
CTO, Model Rocket Feb 2022 – Mar 2023
-
Custom Software Development
Mobile & Web Platform Development & Design
Silver Scintilla was responsive, respectful, and enjoyable to work with, quickly addressing any questions or requests.
DeShawn Brown
CEO & Founder, Lithios Jun 2021 – Nov 2022
-
Custom Software Development
Web-Based HRM System Development
Silver Scintilla provided the best solution with scalable functionality, personalized features, and robust security.
Suheb Khan
Director, Collaborative Insight Technologies Jul 2020 – Jun 2022
Free discovery call
Have an idea? Let's turn it into AI-powered software.
Book a free discovery call with our experts. Share your idea and we will help you shape the scope, timeline and budget, under NDA.
- Free consultation
- NDA before we talk
- Transparent estimate