Methodology
How Sponsor Screening Works
A transparent look at our data sources, methodology, and scoring.
← Back to feature overviewData Sources
Every screening pulls from multiple authoritative data sources. Here's what each provides and how often it's updated.
Companies House API
Company profiles, parent companies and beneficial owners, directors, charges, filing history, disqualified directors register.
Real-time via API
News APIs
National and trade press coverage. Articles classified by AI for relevance, category, and severity.
Checked at time of search + nightly batch monitoring
London Gazette
Insolvency notices, winding-up petitions, and bankruptcy orders.
Checked at time of search
Sanctions Lists
HM Treasury, OFAC (US), and UN consolidated sanctions lists. Fuzzy name matching for identification.
Lists refreshed every 24 hours
Email Verification
Deliverability check, disposable provider detection, data breach exposure count.
Checked at time of search (enhanced screening)
Phone Lookup
Carrier data, number type (mobile/landline/VoIP), disposable number detection.
Checked at time of search (enhanced screening)
WHOIS / RDAP
Domain registration date, registrar, domain age.
Checked at time of search (enhanced screening)
SSL / DNS Checks
SSL certificate presence and issuer, MX record verification, parked domain detection.
Checked at time of search (enhanced screening)
CreditSafe
Company credit scores on connected SPVs in the borrower's portfolio.
Refreshed periodically
Scoring Methodology
Each data source produces individual risk signals. These signals are weighted by severity and combined into a composite score from 0 to 100.
Signal Weighting
Confirmed adverse media in categories like fraud or insolvency carries the highest weight. Softer signals like domain age or email provider type carry lower weight. The score is indicative, not definitive — it highlights where to focus your due diligence, not a replacement for it.
High-Weight Signals
Medium-Weight Signals
Lower-Weight Signals
Score Bands
Low Risk
No significant adverse signals detected. Standard due diligence applies.
Elevated Risk
One or more signals warrant closer investigation before proceeding.
High Risk
Significant adverse findings. Detailed review recommended before any commitment.
AI Classification
Raw news results for common names produce hundreds of irrelevant hits. Our AI layer classifies each result by relevance to the specific individual.
Entity Resolution
Each news result is evaluated against known data about the individual — their Companies House date of birth, company associations, registered addresses, and known roles. The AI determines whether the article is about the specific person being screened or an unrelated namesake.
Category & Severity
Each confirmed or probable match is categorised and scored by severity:
Severity levels: HIGH (confirmed material finding), MEDIUM (regulatory or lower-severity finding), LOW (tangential mention or uncertain match).
Noise Filtering
Only confirmed or probable matches are surfaced in the screening report. Articles about different people with the same name, generic industry news, and irrelevant content are filtered out. This means a screening for “John Smith” returns the handful of relevant results, not the thousands of noise hits you would get from a manual search.
Privacy & Compliance
All data sources are publicly available or accessed via licensed APIs. No personal data is stored beyond what is necessary for the screening report.
Screening reports are stored against your account for audit purposes and can be exported as PDFs. You control your data — screening history can be exported or deleted on request.
Loan Intel is operated by TP Commercial Finance Ltd, registered with the ICO.
Read our full privacy policy →Screen your next borrower in under 30 seconds.
Free tier available. No credit card required.