Methodology
Transparency is a product requirement. Here is how the platform handles facts, estimates and uncertainty.
Model selection
We describe compatibility and trade-offs from disclosed capabilities, deployment options and pricing facts. We do not call a model ‘best’ from vendor marketing or a single benchmark.
Pricing
A price is a dated source fact: dimension, amount, currency, unit, source URL, effective date and verification date. Revisions close the previous record and create a new one. Aggregator-route economics are labelled as such and are not represented as a provider’s direct price.
Cost estimates
The calculator shows its workload inputs and formula. It estimates usage cost from published rates; it excludes provider-specific taxes, minimum commitments, infrastructure, tool calls and unrecorded discounts unless the user explicitly enters them.
Confidence
Future ROI, adoption and productivity figures will distinguish observed, user-entered, estimated and inferred data. High confidence requires direct source data and a complete period; estimates must never be presented as verified financial savings.
Value measurement
AI investment, access, use, workflow change, productivity and economic value will share one data model. Correlation is not causal proof, so comparisons will include baseline, period, source and limitations.