Research / Standards
How Portfilo works with evidence.
We use reporting, documents, data, and other evidence to establish what a story can support. These standards explain how we source information, verify important claims, use AI, correct errors, and keep commercial interests separate from editorial conclusions.
Sources
Start as close to the evidence as possible.
We prefer primary sources when they are available: official datasets, laws and regulations, company filings and records, peer-reviewed research, original datasets, direct interviews, official documentation, transcripts, and records created by the event being reported.
Independent reporting and research can provide corroboration, context, and facts that cannot be reconstructed from primary material. Market platforms, surveys, social posts, and community discussions can also be useful, but we examine what they can actually establish before relying on them.
A person or company saying something proves that they made the claim. It does not automatically prove that the underlying claim is true.
Source principle
Relevance and methodology matter more than a source label alone.
Verification
Important claims have to survive a skeptical check.
Consequential factual claims are checked against their underlying evidence. Before publication, we audit material numbers, dates, names, titles, quotations, company claims, legal or regulatory claims, and other assertions where an error could change the story.
We also look for evidence that challenges the emerging conclusion. Competing datasets, different methodologies, selection bias, weak causal reasoning, stale information, and financially interested sources can all change what a story should say.
If the evidence is incomplete, we qualify the claim, explain the limitation when it matters, continue reporting, or remove the claim. We do not turn missing evidence into false precision.
Unsupported material
- Find evidence
- Qualify the statement
- Remove it
AI use
AI can help with the work. It cannot establish the facts.
AI may assist with
- finding research leads and possible sources;
- organizing and summarizing documents;
- comparing sources and identifying contradictions;
- calculation or code assistance;
- developing questions, drafts, and edits after the writing form is chosen;
- skeptical review that points to claims needing another check.
AI cannot substitute for
- a source;
- evidence;
- verification;
- reporting;
- editorial authority for structure, voice, or institutional claims;
- authority for a quotation, statistic, attribution, law, or current fact.
Factual material surfaced by AI must be independently checked before it can be published. AI output is not treated as authority for how a story should be structured or how Portfilo should describe itself.
Corrections
Material factual errors are corrected visibly.
When a consequential factual error is found, we correct the story and attach a visible correction note. The note should state the correction date, what was wrong, and what is correct.
We do not silently rewrite a material error in a way that hides the correction. Minor typographical or formatting fixes do not require a formal correction unless they change meaning.
Correction record
- Date
- When the correction was made
- Wrong
- What the story previously said
- Correct
- What the evidence supports
Commercial independence
A commercial relationship cannot determine the factual conclusion.
Sponsors, advertisers, partners, and affiliates may not determine the factual conclusion of Portfilo editorial coverage.
Commercial material must be clearly labeled. Advertising should not be presented as independent reporting. If a commercial relationship is relevant to how a reader would interpret a story, it should be disclosed.
Editorial rule
Commercial support does not buy an editorial conclusion.