Every SMS a carrier handles is classified as one of two kinds of traffic, and which one yours is decides how it is routed, priced and filtered. The classification is about who is sending rather than what the message says.
What is the difference?
P2P, person to person, is a conversation: one individual texting another from a handset, typically balanced in both directions and modest in volume.
A2P, application to person, is a business, organisation or application sending to individuals. Passcodes, delivery notifications, appointment reminders and campaigns are all A2P, whether or not a person composed the text.
The distinction is not about tone. A friendly, personalised message sent by your application is A2P. A commercially significant message typed by a human on a phone is P2P.
Why do carriers care?
Because the two behave differently, and one of them is where abuse lives.
Industry guidance draws the line explicitly. The CTIA messaging principles treat consumer and non-consumer messaging as separate categories with different expectations, and describe typical consumer behaviour in terms a filter can act on: roughly balanced traffic in each direction, unique rather than repeated content, and messages originated by a person.
A campaign has none of those properties. It is one-directional, near-identical across thousands of recipients, and machine-originated. That signature is easy to recognise, which is why A2P traffic is expected to arrive on routes and sender types intended for it, with the sender registered.
What is a grey route?
Sending A2P traffic over a path that was not authorised for it, and it is a bad idea rather than a clever one.
The CTIA principles name grey routes as something message senders should not use, defining one as a setting, method or path a service provider has not authorised for non-consumer messages. The appeal is obvious: consumer routes are cheaper and skip registration. The cost is that the traffic is unregistered and unexpected, so it is filtered aggressively, blocked without notice, and gives you no route to complain.
The related pattern is snowshoeing: spreading a campaign across many sending numbers so each one looks low-volume and conversational. The same guidance names it explicitly as something senders should not do. Both are attempts to make A2P traffic pass as P2P, and both are recognised failure modes rather than undiscovered loopholes.
If your messages are being filtered without an obvious cause, the route and the sender are worth checking before the content. Why are my SMS messages being filtered by carriers covers what the failure actually looks like from the API.
What does this mean for my setup?
Mostly that registration is not optional paperwork, it is how your traffic gets classified correctly.
Registration is the mechanism by which a carrier knows who is sending A2P traffic and what it is for. The programme that does it in the United States is 10DLC, and other markets run their own sender-ID pre-registration; what each one demands is published by the body that runs it rather than summarised here. What is sender ID registration covers what a registration records, Bird's 10DLC guide covers the US programme, and which sender type should I use covers picking one.
The practical version: choose a sender type intended for business traffic, register it where the destination requires registration, and send from it. That is the whole of behaving like A2P traffic, and it is considerably cheaper than the alternatives.