below are my thoughts on AMP’s taxi.com.br idea…
.
It can become a mix of vans (as in more than one passenger) and cabs (passenger defines point A and B), using networks to calculate the most optimal source. After all, fixed stops & fixed schedules were needed before the age of constant communication.
.

.
Check seamlessweb.com business plan: Customers use the web to order deliveries (credit card payment, better menu list, convenience) and restaurants don’t need to lock a line and pay employees to be available to call. It sits in the middle, helps both parties and gets a cut of the action.
.
Same way with your taxi.com.br… I’ll send you a paper about it if j can find it, but here’s the happy path:

• Customer goes online (minimal website interface, good but not limited to iPhones) tells point A (where he is), point B (where he wants to be) and optionally, when (defaults to as soon as possible)
• customer receives total price, pays right away
• customer receives an sms telling the 5 min before van arrives (the driver has to inform the system, more on that later)
• customer picks the van, is told when it arrives to point B
• customer receives an email asking to rate the experience.

Price fluctuates based on van availability on the area a or b, traffic, reputation.
.
The van is part of a club. It needs:

• initial checkup
• app installation fee (you sell)
• monthly fee (comm costs)

The app can be a simple beep with gps.
.
The higher the ratings:

• lower the monthly fee
• better offers (daily commute pool, fixed A, B points)

The driver doesn’t have to negotiate payment or service. The system deals with it and offers customers from the nearest places (route optimization). The driver gets the fare, a cut going to the system (70-30?)
.

Problems:

• legislation, lobbies
• quality of service
• do we count start or end time? End time is better, but we need to take traffic into account

Opportunities:

• airports, malls, any fixed point b
• daily commutes: customer pays monthly and can make minor adjustments (free prior 24h). You’re competing with cars!!!
• if overall cost is cheaper than car, with same benefits (control of a to b), market it as the green solution, find government grants