Jovian AI’s evidence, scoring, disclosure and correction method for future product recommendations.
A useful recommendation should explain what is known, what remains uncertain and why one option fits a defined use case. Jovian AI does not treat popularity, price or commission rate as sufficient evidence of suitability. This article describes the methodology intended for future recommendations.
Four evidence labels
Every material statement should belong to one of four categories. A sourced fact is supported by an identifiable, dated source. A direct observation is recorded during a test Jovian AI actually performed. An estimate is calculated or projected from stated assumptions. An editorial opinion is a judgment based on declared criteria. Manufacturer and retailer statements remain sourced claims until independently observed. We do not describe a product as hands-on tested when it was not.
Define the use case first
Before comparing products, we define the intended user, required task, physical or technical constraints, acceptable setup complexity, budget boundary and factors that would disqualify a candidate. This prevents a general feature list from being mistaken for a recommendation.
Score decision-relevant criteria
Criteria may include functional fit, reliability evidence, compatibility, setup friction, portability, maintainability, total expected cost, rights feasibility for media and time to useful deployment. Weights must be stated before final ranking. Missing evidence should reduce confidence rather than be silently treated as favourable. A high score does not override a failed safety, rights or compatibility gate.
Separate research from direct testing
A research-based comparison can help narrow a category, but it is not a substitute for physical testing. When direct testing is unavailable, the article must say so and focus on specifications, decision criteria and unresolved questions. If testing later occurs, the article should describe the setup, conditions and limitations instead of generalizing from one trial.
Keep commercial relationships visible
Affiliate eligibility or commission does not determine whether a product qualifies for consideration or where it ranks. When affiliate links are present, disclosure must appear clearly before or near them. At initial publication, this article contains no affiliate links.
Correct and update
Each recommendation should display an author, publication date and updated date. Material corrections should be made transparently. Variable facts such as availability, specifications and program terms should be rechecked before reuse.
Evidence status
Facts: the published methodology and article metadata. Direct observations: none claimed in this methodology draft. Estimates: any future commercial or performance forecast, with assumptions. Editorial opinions: rankings and fit judgments supported by disclosed criteria.
Future product articles should cite primary documentation for specifications and identify independent sources where they materially support a claim.
Original Jovian AI editorial methodology. Future product claims require identifiable, dated sources; no affiliate link is present.