Hawala (sometimes referred to as hundi) involves a trust relationship between the two hawaldars. The hawalder who receives the funds (in Minneapolis, for example) is a trusted partner to the other hawalder who disburses the funds (in Somalia, for example).
With Bitcoin, this relationship between hawalders becomes unnecessary. These hawalders simply become exchangers operating independently.
There are plenty of methods in which bitcoins can be acquired by the person wishing to send money, so the remaining gap is finding an exchanger who will trade shillings, dollars, euro or whatever for bitcoins within Somalia, for example.
Bitcoin is uniquely positioned to make this possible however, due to its properties where transactions are non-reversible, extremely low in cost (transaction fees), and where funds become available for spending right away (in about an hour, for example).
This makes it possible for a network of individuals to do the grunt work function where the exchange involves transaction sizes as low as $20 worth of bitcoins, for example. With transaction fees being trivially low, small, independent exchangers can trade bitcoins with each other for immediate needs but when the supply of shillings, dollars and euro needs to be replenished they can trade with slightly larger exchanges.
Those larger exchanges would then trade bitcoins either directly or indirectly with Bitcoin exchanges that interact with banks or with parties that essentially offer these banking services.
When bitcoins then are seen as something with value because they can easily and inexpensively be cashed out, on demand, they might then start being used for trade alongside shillings, dollars and euro.
And that is where things start to become interesting.