Reliability is the quality travellers notice most and compare least. Price is one number and easy to sort. “Will this flight leave roughly when it says?” takes work to answer, so most people don’t.
It is partly answerable — with an important limit that’s usually left out.
Where the data comes from
In some markets, carriers are required to report operational performance publicly, including on-time arrival rates by route and by airline. In the United States this reporting is unusually detailed, which is why US domestic routes can be assessed with real evidence rather than reputation.
That’s genuinely useful. A route’s historical on-time rate is a far better guide than an airline’s general reputation, because performance varies enormously within a single carrier depending on the route, the aircraft rotation and the hub.
The limit nobody mentions
Coverage is uneven, and the gaps are not random. Public reporting requirements apply to certain carriers in certain markets. Many international routes and many smaller carriers are not covered at the same depth or at all.
This creates a trap. If a tool shows an evidence-based reliability score for a covered flight and a default placeholder for an uncovered one, and displays them identically, you are not comparing two measured things. One side is measured and the other is assumed — and the assumed one is often the foreign carrier, which is exactly where a traveller most wants real information.
Where data is absent, the honest presentation is to say so, not to fill the gap with a neutral-looking number that reads like a measurement.
| Flight | What the figure is | How it should be shown |
|---|---|---|
| Covered by public reporting, such as many US domestic routes | Measured: a historical on-time rate for the route | As a measured figure |
| Covered thinly or not at all, such as many international routes and smaller carriers | Little or nothing measured; any number would be assumed | As missing, not as a neutral-looking placeholder |
What you can do
- Prefer earlier departures when it matters. Delays accumulate through the day as aircraft fall behind their rotations.
- Treat the first flight of the day as the safest slot, since its aircraft is usually already in position.
- Check whether the figure you’re shown is measured or assumed. If a tool won’t tell you, treat the number as decoration.
BookWithAI is explicit about this: coverage is strongest on US domestic routes, and on-time data is limited outside the US. You can read what BookWithAI can and cannot tell you about a route, including the limits stated plainly rather than buried.