Confidence, with the working attached.
Pretira supplies valuation, attribute and risk data for UK lenders and banks - built so modelled outputs can be reviewed alongside available confidence, provenance and test evidence, with the scope stated explicitly.
The confidence level is the product.
A modelled valuation without an indication of confidence is hard to route safely. Pretira is designed to present the estimate, its available confidence indicator and a sensible range together, so lenders can define the review rules that fit their policy.
Declared and modelled, separated.
We publish which fields come from record evidence and which are modelled, and we test the models against the declared set. Nothing is blended silently, so your second line can weight the evidence rather than take it on faith.
Sold prices are the benchmark.
Completed transactions provide a useful benchmark for valuation performance. Where applicable, an evaluation can compare predictions with published completions. The relevant cohort, test period, metrics and recalibration approach must be stated for the specific product rather than implied by a universal claim.
The working is available for review.
Relevant source attribution, methodology for model-risk review, a data dictionary and a security review pack can be provided according to the selected product and evaluation scope. If your process needs a DSA, a DDQ and committee review, we can plan those requirements into the evaluation.
The lender-facing product brings together data, methodology documentation and an evaluation process. Available confidence, provenance and test evidence are provided at the level supported by the selected output; source and third-party dependencies remain identified.
Bring the cases your current data declines.
Thirty minutes, no slides. We look at them live - and you see the working, not just the number.