Yep, of course the ISP cells/towers are omnidirectional and roughly cover a circular area around them with “good” signal.
If you are within that area or close to it having an external antenna is not really needed.
On the other hand, if you are outside that area, you might need an external antenna (not necessarily directional).
If you are far from the cell/tower the only way to have decent signal is an external directional antenna (and some luck).
In real world things are not as simple as that, as there may be several factors influencing the reception.
To give you some real world anecdata, I have a FWA (LTE/4G) connection at home which more or less sucks.
But it sucks less then it used to, since the antenna modem was changed with a “better” one (imagine LTE 12 replacing LTE6).
Then, I have very marginal coverage from two cell towers (so I am in one of those cases where I need a directional antenna), one which is nearer and stands in a “perfect position” on top of a hill, the other one is roughly at the same elevation as my house and about 1 km further.
When the antenna/modem was replaced it was initially oriented as the old one to the nearest tower, and still sucked, then the technician tried the other tower and suddenly there was some clearly much better bandwidth/speed (and additionally, as we could experience in the following weeks the connection was much more stable).
No apparent reason if not - maybe - that the furthest tower was installed more recently and thus had “better/faster” devices than the nearer one.
It is clear that real life speed is a fraction of the theoretical one, in first (very rough) approximation I would say 1/30 to 1/10, possibly much less depending on local conditions.
Check this post for a couple of well explained examples:
http://forum.mikrotik.com/t/lhggr-underperforming-lte-speeds/176000/1