A multitude of actual answers were developed during the period when ORION nuclear pulse drive were being actively investigated, but references are scattered all over the place, making it difficult to encapsulate this. Some of the best sources of information are through the Atomic Rockets website and the "Unwanted Blog", who's author is a big fan. A man named Anthony Zuppero also came up with an.....interesting......use for ORION, which will also be discussed.
While 4000 ton space battleships were seriously being proposed (having Minuteman ICBM's as their main battery, 5" naval gun turrets for close defence and so on, this was really more than a bridge too far. Much smaller ORIONs were pitched to NASA and the USAF, with the size constraint being the diameter of the drive plate had to fit a Saturn V booster for lift off.

Model of the 4000 ton space battleship proposal. Yes, this was actually a real proposal

USAF/NASA "10 metre" ORION. This was to be boosted by a Saturn V into orbit. From the Unwanted Blog
By 1965, the ORION program was winding down, but General Atomics tried to sell the USAF on the program with a scaled down set of spacecraft, such as an emergency command and control center (blasted into orbit from a ground silo), an orbiting command post and a "bomber" on a long looping orbit which took it past the Moon and made it extremely difficult to intercept or eliminate America's nuclear deterrent. These were meant to be lofted by clusters of solid fuel boosters ranging from 7 120" solid fuel boosters (derived from the Titan III boosters) to 156" solid fuel boosters (4 for a lightweight ship and 7 for a heavy ship...)

Emergency command post lifting off

Model of proposed USAF ORION command post spacecraft

Bomber ORION on it's launch stack
So ORION could be launched by various means. The launching of ORION using rockets actually invalidated much of the reason for having ORION in the first place: being able to launch heavyweight spacecraft from Earth. The idea of radioactive fallout and EMP from launch made launching ORION using nuclear pulse units a political non starter. However, if it is the end of the world, then perhaps that is no longer a consideration.
Anthony Zuppero wrote an interesting memoire about his time working in the defence industry. as a thought experiment, he was asked to find a way to disable Soviet nuclear missiles in their launch fields "in two minutes" in order to prevent them from being launched. Zuppero went on to design perhaps the largest semi plausible weapon ever - a 5 gigaton weapon launched from the United States on an ORION pulse drive...The thought of the largest nuclear weapon ever conceived riding skyward on a pillar of nuclear fire is, well...

Schematic of the Doomsday ORION by [William Black]. The design for the actual 5 Gigaton warhead is strictly conjectural.11
So in terms of making ORION a plausible vehicle, there is no doubt that it was, and using 1960 era technology as well. The idea of a high thrust, high ISP drive is still viable, but would have to be only for very extreme conditions, such as saving the world:
http://nextbigfuture.com/2009/02/unmanned-sprint-start-for-nuclear-orion.html
An unmanned Orion asteroid interceptor was designed. It would not need shock absorbers. Artillery arming, fusing, firing system for shells are regularly built to take 1000 Gs.
There was a three page paper: Nuclear explosive propelled Interceptor for deflecting objects on collision course with Earth. Johndale Solem, Los Alamos, proposed unmanned vehicle. No shock absorber or shielding. The pulse units were 25kg bombs of 2.5 kiloton yield.
Get to high velocities with only a few explosives and small shock absorbers or no shocks at all. Launch against a 100 meter chondritic asteroid coming at 25 km/sec. 1000 megatons if it hits. Launch when it is 15 million kilometres away and try to cause 10000km deflection. A minimal Orion weighing 3.3 tons with no warhead would do the job. 115 charges with a total of 288 kiloton yield. Launch to intercept in 5 hours. Ample time to launch a second if the first failed.
Outside of asteroid destroying interceptors, the ability to use ORION to launch heavy payloads from the ground is pretty much a dead letter. However, since ORION nuclear pulse drive is the only plausible high thrust, high ISP drive currently known (fusion drives need to actually demonstrate fusion before they can be considered), it will be highly desirable for deep space missions. In order to get the maximum performance from ORION, and use the smallest possible pulse units (people will be nervous about nuclear detonations in space, despite the fact space is already a high radiation environment, and particles from a nuclear explosion are moving at far beyond escape velocity, so will never return to Earth), the ORION formula will have to be inverted.
MEDUSA is a proposal to "invert" the ORION, with the plasma from the pulse projected against a gigantic "sail" ahead of the spacecraft. The thrust was transmitted to the spacecraft via a huge network of cables, making the entire structure in tension, and thus far lighter than the equivalent pusher plate structure. Since the sail is far larger than any conceivable pusher plate, much more of the energy can be captured. Combining this with a much lighter overall spacecraft system makes MEDUSA a wonderful method of transporting large payloads to the outer Solar System in a fraction of the time compared to other proposed methods.

Schematic of MEDUSA in operation

MEDUSA nearing Saturn
So the best way to make ORION a feasible and workable system for actual use is to have it as a deep space propulsion system, capable of moving large payloads to the outer Solar System in reasonable amounts of time compared to ion, nuclear thermal or other known systems. This in turn requires some plausible and economically feasible reason to actually go to the outer Solar System. A few plausible reasons may exist:
Anti protons are captured by the magnetospheres of planets. The giant planets, with their huge magnetospheres, will capture usable amounts (micrograms, but antimatter is extremely potent), making rapid deep space travel economically viable. Capturing antimatter and sending it to Earth as an extremely compact energy storage and generation method is a plausible reason to do deep space missions.
Titan's atmosphere contains millions of tons of hydrocarbons. These are useful as chemical feedstocks for plastics, pharmaceuticals and fertilizers, among other things. A space economy will need these products, and having a "local" source that does not need to be lifted from a deep gravity well is a bonus.
Titan's atmosphere and extreme cold conditions also make it potentially the industrial powerhouse of the Solar System. The Carnot equation tells us efficiency is measured by the differential between the hot and cold side of a system. On Earth, the "cold" side is at atmospheric temperatures, but on Titan, the atmospheric temperature is hundreds of degrees below 0 C, meaning most processes and computing become far more efficient compared to the same things on Earth. This makes cheap, high speed transport vital.

Carnot Efficiency. When the difference between tH and tL are larger, the efficiency goes up
So with sufficient reason to actually go to deep space destinations, ORION and its variants will be highly desirable due to their high thrust and high ISP, making fast and cheap transportation possible.