Corrections and Updates Are Different
We distinguish between a correction and an update. A correction fixes something Credit Consensus Index reported incorrectly. An update reflects information that changed after the original publication.
A correction may involve:
- Incorrect fee or APR range
- Incorrect recommendation count
- Misidentified company or product
- Calculation, data-entry, or attribution error
- Wrong eligibility requirement or product feature
- Misquotation or misinterpretation of underlying research
An update may involve:
- A lender changing its APR range
- A credit card changing its annual fee
- A provider changing pricing, features, or eligibility
- A product being discontinued
- A new research cycle producing different AI recommendations
If the original information was accurate when published and the company later changed its terms, that is an update, not an error.
We Correct Material Errors
When we identify a material factual error, we aim to correct it. Material errors are facts that could reasonably affect a consumer decision, ranking, recommendation, product comparison, eligibility, cost, risk, or interpretation of our methodology.
- Reporting the wrong annual fee, loan range, or security deposit
- Misstating bureau reporting or a credit product feature
- Miscounting AI platform recommendations
- Assigning a recommendation to the wrong company
- Describing a provider as available in a state where it is not
Minor typographical errors, formatting problems, broken links, misspellings, grammar issues, and non-material wording improvements may be corrected without a formal notice. A public correction note is more appropriate when the error could materially affect a reader’s understanding.
Material Corrections Should Be Transparent
When a meaningful factual error is corrected, we may explain what was incorrect, what was corrected, when the correction was made, and whether the ranking or conclusion changed. If a correction changes a ranking, the article should reflect the corrected ranking and, where practical, disclose that effect.
We do not preserve an incorrect result merely because the article has already been published. Accuracy is an editorial responsibility, and legitimate factual corrections are not conditioned on advertising, affiliate status, consulting, data purchases, or payment.
Historical Research Should Remain Historical
Credit Consensus Index studies AI recommendation behavior at a specific point in time. If a company ranked first in January and fourth in July, July’s result should not erase January’s result. Where practical, we preserve the original research date, historical ranking, recommendation coverage, previous study version, current ranking, and current coverage.
A new research cycle is not normally a correction. If a company was recommended by 6 of 7 platforms in January and 3 of 7 in July, the January study was not necessarily wrong; the recommendation environment changed.
We do not rewrite raw AI responses to make an AI platform appear more accurate. We preserve what the system actually said and may add a verification note explaining that current provider disclosures differ. Product changes, acquisitions, rebrands, affiliate changes, and model renames may be reflected in current information while preserving enough historical context to identify what was tested.
Research Dates and Product Updates
Research Date identifies when underlying AI responses were collected. Published Date identifies when an article was first published. Last Updated identifies when meaningful content changed. Reviewed Date identifies when a human reviewer performed a substantive review. These dates are not interchangeable and should not change merely because a comma, typo, link format, or layout changed.
We may update a current fee or product term without rerunning the AI study. When we do, the article should make clear that the product information changed while the historical AI research date remains the same. We should not mix research cycles without disclosure.
Financial information requiring ongoing review includes APRs, fees, loan amounts, terms, prequalification, state availability, security deposits, credit limits, rewards, credit reporting, subscription pricing, debt minimums, cancellation terms, and program structure. Readers should verify material terms directly with the provider.
What May Trigger an Update
Updates may be triggered by:
- New AI research or methodology changes
- Product, fee, APR, eligibility, or service changes
- New regulatory information or material risks
- Product discontinuation, acquisition, or rebrand
- Correction submissions or internal quality-control review
- Major litigation, enforcement, bankruptcy, licensing problems, breach, or consumer-protection concern
High-impact information receives priority, but we cannot guarantee instant updates. Companies and AI platforms may change without notice, so there may be periods when published information is no longer current.
Companies and Readers May Request Corrections
Companies discussed in our research and readers may report suspected errors in good faith. A useful submission should identify the exact statement, page URL, why it is incorrect, the proposed correction, supporting evidence, and a way to ask follow-up questions.
- Broken links or outdated pricing
- Incorrect features, calculations, or state availability
- Misidentified credit score models or bureau reporting
- Incorrect platform or recommendation counts
- Official product disclosures, current provider terms, regulatory records, government sources, court records, or company documentation
A company does not automatically receive a correction merely because it disputes our characterization, ranking, AI recommendations, or interpretation. A branding preference is not necessarily a factual correction, and we do not manually add AI consensus that did not occur.
Submit a request through the Contact Us page.
Our Editorial Process
We evaluate evidence independently of commercial relationships. Corrections should preserve context: changing one number may require updating surrounding language about cost, risk, or suitability. If a correction affects consumer risk, ranking, product suitability, conclusions, legal context, or material cost, the article may receive a new substantive human review.
We may remove, qualify, attribute, or flag material information that cannot be verified after reasonable review. In unusual situations, a notice may state that information is under review. We may also consolidate duplicate or obsolete pages while preserving important research meaning and distinguishing historical findings from current ones.
We maintain research records where technically practical, including prompts, platforms, responses, research dates, recommendation extraction, entity normalization, and ranking calculations. Auditability helps readers, companies, editors, reviewers, and researchers investigate discrepancies.
Corrections to AI Platform Names or Models
AI companies frequently rename models, products, modes, and interfaces. We may update current platform terminology while preserving enough historical context to identify what environment was actually tested.
Research Records Should Be Auditable
Where technically practical, Credit Consensus Index should maintain internal records sufficient to review the prompt, platform, response, research date, recommendation extraction, entity normalization, and ranking calculation. These records help us investigate correction requests and data discrepancies.
Preserving research records helps readers, companies, editors, reviewers, and researchers. If someone challenges a ranking, we should be able to return to the underlying data rather than rely solely on memory.
What We Correct, Update, and Preserve
What we correct: our own mistakes.
What we update: facts that have changed.
What we preserve: accurate historical research.
That is the foundation of this policy. Before closing a material correction review, we aim to ask: Was the original statement actually wrong? Was it accurate at publication but later changed? Does the supporting source establish the correction? Does the change affect the ranking or conclusion? Should the article receive a correction notice? Should the reviewer examine the updated content? Can the historical AI research remain intact?
These questions help us distinguish correction from revision.
Why This Policy Matters
Credit Consensus Index is intended to become a historical record of how AI systems recommend financial companies and products over time. That record loses value if we constantly overwrite the past. But consumer financial information also becomes dangerous if we refuse to correct it when it changes.
Our responsibility is therefore twofold: preserve the integrity of historical research and maintain the accuracy of current consumer information. Both matter.
The Principle Behind Our Corrections Policy
Our corrections philosophy can be summarized in one sentence: Correct our mistakes, update what changed, and preserve an accurate record of what the AI systems actually said.
Important Information
Credit Consensus Index makes reasonable efforts to keep material consumer-credit information accurate and current, but financial products and AI systems can change quickly. Verify rates, fees, eligibility, availability, deposits, and other material terms directly with the provider before making a financial decision.
A correction to current product information does not necessarily mean the underlying historical AI research was inaccurate. Compensation from companies discussed on this website does not affect whether legitimate errors are corrected or accurate historical research remains published.
Product Discontinuation, Acquisitions, and Rebrands
When a product is discontinued, we may update the article, mark the product as discontinued, preserve historical ranking information, remove current application links, or identify replacement products where appropriate. A discontinued product's historical AI ranking may still be useful research.
Companies may be acquired, merge, change names, or move products to a new brand. We may update current entity information while preserving enough historical context to explain what the AI systems originally recommended. AI companies may also rename models, products, modes, or interfaces; we may update current terminology while preserving enough context to identify what environment was tested.
Commercial Relationship Changes
If a company becomes or stops being an affiliate partner, changes commission, or ends its program, the underlying research does not change. We may update links and disclosures, replace broken affiliate links with a direct provider link, another valid tracking link, or a non-commercial link, without changing the company's ranking.
If a sponsorship begins after publication, we should disclose the relationship where relevant. If it ends, we may update the disclosure. Commercial changes do not retroactively alter independent research results.
Historical Archives and Methodology Changes
Where practical, older research versions or ranking history may remain accessible and may be labeled archived research, previous ranking, superseded study, historical data, or earlier research cycle. We may consolidate duplicate or obsolete pages, redirect URLs, archive duplicate studies, or update navigation while preserving important research meaning.
Methodology changes may include a different platform set, ranking calculation, qualification threshold, entity normalization, research tooling, or source-verification procedure. Material changes should be documented. New and old studies may not be directly comparable, and we do not necessarily retroactively apply every new rule to historical research.
Update Frequency and Priority
Different information requires different review schedules. Priority may depend on consumer impact, traffic, product volatility, research importance, known changes, and regulatory developments. High-impact information such as APRs, fees, eligibility, product shutdowns, state availability, deposits, debt minimums, major regulatory actions, and credit reporting receives priority.
We cannot guarantee instant updates. Financial companies may change terms without notifying us, and AI platforms may change behavior without advance notice. Readers should verify important current information directly with the provider.
Our Corrections Standard
Before closing a material correction review, we aim to ask: Was the original statement wrong? Was it accurate at publication but later changed? Does the evidence establish the correction? Does it affect the ranking or conclusion? Should the article receive a correction notice or a new human review? Can the historical AI research remain intact?
We do not charge for legitimate corrections, require advertising, affiliate status, consulting, data purchases, or payment, or accept payment to remove accurate unfavorable research. We may decline requests to remove historical AI findings when the dataset accurately records what an AI system said.