Iro AINamed authorship
There is no content team and no freelancer pool. Every article and topic page on this site is written by Alex Furukawa, the founder of Iro AI. He leads the AI initiative at a private equity real estate firm and builds the automation his team uses on live deals — so when a page describes a workflow, it is one he runs, not one he read about.
That is a deliberate constraint. It means less content than a site with a writer pool, and it means the byline is accountable to a real person. Author entity: https://tryiro.com/#author. Bio on the About page.
Product claims trace to one source
Every figure about Iro — 29 paths, 477 lessons, 3,000+ exercises, 24 exercise types, pricing, and exactly what is free versus Pro — is checked against the app's own data before it goes on a page. The machine-readable reference at /llms-full.txt is updated in the same pass.
If a page and that file disagree, the file is canonical. That rule exists because the failure mode of a growing site is not lying, it is drift: a number that was right in June and was never revisited. We run an automated check before publishing that flags any page contradicting the current product facts.
Third-party figures carry a date
Competitor prices, App Store ratings and refund terms change without notice, so we publish them as US figures at the time of writing with a last-reviewed date attached to the table they sit in.
Where we have not verified a number ourselves, the cell says Not verified by us rather than repeating a figure we found elsewhere. A confidently stated wrong price is worse than an admitted gap.
Claims are separated by certainty
Not everything on a comparison page is the same kind of fact, so we do not present it that way:
- A competitor's published policy is quoted and attributed to them.
- User reports and review-site complaints are labelled as user accounts, at a lower level of certainty, and never upgraded into established fact.
- Our own opinion is written as opinion — "we think", "the honest test is" — not disguised as a finding.
Comparisons disclose the conflict
Iro appears in several of our own roundups, including the best AI learning apps and the free ones. Every one of those pages states plainly that Iro is our app, and holds it to the same criteria as everything else in the table — including the rows where a competitor is the better answer. A roundup that hides its own conflict is marketing wearing a comparison's clothes.
Original data ships with its limitations
Our AI literacy figures come from our own free AI IQ test. They are published with the sample size, the collection window, and the caveats that weaken them: respondents are self-selected, retakes are included, and the founder's own attempts are disclosed rather than quietly excluded.
Small original data with its limits stated is more useful than a big number with none.
Corrections and changes
Errors of fact are corrected on the page itself and the modified date is updated — we do not silently delete a claim and move on. Product changes are logged in the changelog, so a reader can see when something they read stopped being true.
If you find something wrong on this site, the contact channels are on the About page. Corrections are the cheapest trust we can buy.