Regression and gradient boosting models (XGBoost or LightGBM) trained on comparable sales data in your target market. Features include: bedrooms, bathrooms, floor area, plot size, property age, condition rating, EPC rating, distance to nearest school (Ofsted Outstanding), distance to nearest station, crime rate by postcode, flood risk zone, and local sold price index per quarter. Produces property value estimates with confidence intervals in seconds -- a point estimate plus a 90% confidence range based on the variance in comparable transactions.

Model validation runs against a holdout set of recent sales before deployment: Mean Absolute Percentage Error (MAPE) and the percentage of estimates within 5% and 10% of actual sale price are the primary accuracy metrics, reported per property type and price band so you know where the model is reliable and where it is less certain. Applicable to portfolio tracking, initial vendor guidance (agents using it as a starting point for valuation conversations, not as a replacement for a surveyor), mortgage underwriting support (pre-screening before instructing a full survey), and acquisition screening at volume where manual comparables analysis is not feasible. Explainability output shows which comparable transactions and attributes drove the estimate -- so an agent can discuss the basis of the figure with a vendor rather than presenting a black-box number.

Classification models trained on your historical enquiry and conversion data. Scores each new lead at intake by conversion likelihood: which enquiries are most likely to result in a viewing, offer, or letting. High-score leads surface to your agents for immediate follow-up. Low-score leads enter automated nurture sequences. Reduces agent time spent on cold leads and improves response time for high-intent enquiries. Requires 6-12 months of enquiry history with known outcomes to train.

Semantic search and matching models that go beyond keyword and filter-based search. Understands natural language queries: a buyer who says "quiet street, near good schools, south-facing garden" gets results matched against those attributes rather than a keyword filter returning nothing. Matches buyers and tenants to properties based on stated and inferred preferences from engagement behaviour. Applicable to both external search experiences and internal agent tools for matching buyers to new listings.

Document extraction pipelines that read lease and contract documents and extract structured data fields: tenant details, dates (commencement, expiry, rent review dates, break option dates, notice periods), current rent, rent review mechanism (RPI-linked, fixed uplift, open market review), break clauses (conditions, notice requirements), permitted use, repair obligations, alienation provisions, and special conditions. Handles PDFs (digital and scanned), Word documents, and varied template formats from different solicitors and time periods. OCR preprocessing with layout analysis preserves table structures and clause boundaries that a plain OCR pass loses.

Output is searchable and reportable structured data: a commercial property manager can query "all leases with a break option in the next 18 months" or "all tenancies where the rent review mechanism is open market and is due within 6 months" without opening a single document. The extracted lease data integrates with your property management system (Yardi, MRI Software, Re-Leased, or custom) via API. Makes your lease portfolio queryable without reading each document. Essential for commercial property managers and landlords with large portfolios where manual data extraction is not operationally viable -- a 200-property portfolio with an average 30-page lease represents 6,000 pages of unstructured data this pipeline converts into a usable database in days rather than the months manual extraction would require.

Models that recommend rental asking prices based on current market demand signals, comparable listing and transaction data, seasonal patterns, and vacancy risk. Responds to current market conditions rather than historical comparables. Recommends when to adjust price for listings generating insufficient enquiry volume. Helps portfolio managers and lettings agents minimise days-to-let while maintaining rental income. Trained on local market data combined with your historical listing performance.

Tenant churn prediction models that score renewal risk for each tenancy using payment history, maintenance request frequency, engagement with renewal communications, and lease expiry proximity. Surfaces at-risk tenants for proactive outreach before the decision to leave is made. AI chatbots for property enquiries that handle initial qualification questions, book viewings, answer frequently asked property questions, and collect contact details outside business hours. Both systems connect to your CRM via API.