People program up FMGC and try to fly a visual through the automatics and map display.
Up to the point where I've seen guys try to program the visual on 31 in LaGuardia - not a good idea.
Apropos another thread, we do not permit night visuals in designated mountainous terrain for a good reason.
That said, a well-flown visual from a mile or two outside the marker, (4 to 6 miles back), when in the slot can save fuel and time but I agree with Jaxon - the ability to fly a visual approach simply by looking outside is a disappearing art. Disconnecting the a/p and a/t below 10 and hand-flying a visual is a good way to get everyone's close attention in the cockpit, first because many have never seen it done...
With elementary use (not reliant-use) of the map feature one can maintain situational awareness while employing the one-in-three rule for altitude and using the groundspeed-divided-by-2 method that yields the rate-of-descent to maintain for a rough 3-deg visual "g/s" all the way down, keeping in mind how much distance it takes to slow down and get configured. It's simple and everyone who's spent any time hand-flying (when appropriate), and not merely "managing" the airplane all the time knows this like the back of their hand but, along with the go-around maneuver, a visual approach can be one of the most cocked-up maneuvers seen, even in the data.
On a matter related to risk, I have heard it said that the go-around maneuver may not always be the preferred "out" to an approach that has become destabilized, (not necessarily referencing the Air France 358 TSB Report here). To me that's nonsense because the maneuver is (or ought to be) basic toolkit stuff, but the observation was made and deserved some thought about accidents related to the go-around maneuver. The context was, whether go-arounds ought to always be encouraged when an approach becomes destabilized. I couldn't really believe my ears but there it was.
One that comes to mind both in the go-around and then poorly-executed visual approach(es) was the Gulfair A320 accident at Bahrain in August, 2000. There have been several "mode-confusion" go-around accidents since then which may have had a vestibular disorientation component. Training issues emerge as well.
Perhaps this is for another thread, (admin?, & if there is indeed any interest in the question), but what sense is there that a go-around can approach the riskiness of continuing a highly unstable approach?