It depends entirely on how desperate the need is
Or rather, what likelihood of failure the organisers are willing to accept.
We already have a design for a machine that can take three people to the moon, allow them to stay there for 24 hours and return. We went from zero to first launch of the Saturn V in about 5 years. And the plans still exist, and are even public domain. By far the simplest way to get back to the moon is to take them off the shelf, knock off the dust and start building.
Except that that's crazy, because that technology is old, and that's a problem for two reasons. Firstly the supply chains (indeed the entire industries) that support the availability of the component parts are gone and would also need to be rebuilt, an exponentially expanding logistical challenge. And secondly, the capabilities of the equipment are, by modern standards, horrifically limited and unsafe. The Saturn V autopilot, for instance, was a computer with a clock speed of 2.048MHz and a processing capability of 12 kiloflops. That's less powerful than an Arduino. The main computer of the Apollo command module was the size of a suitcase (and would incur excess baggage charges by most airlines) and not significantly more powerful. Components were huge, weighty, very limited in their functionality and would be oh so easy to replace with modern components which would be much easier to manufacture with modern supply chains and far more feature-full.
Assuming that they work correctly. Assuming that replacing a 30kg lump of metal with a single featherweight integrated circuit won't skew the centre of gravity of the ship in an unexpected way. Assuming that swapping in 3D-printed titanium or carbon fibre instead of milled lithium or whatever else they used doesn't have some unexpected effect. Assuming that whatever interface they introduce to interface 32- or 64-bit modern electronics to the 14- or 15-bit original electronics doesn't have any unexpected bugs. And so on and so forth.
You can iron out all these bugs with testing, the more extensive (and lengthy) the better. The longer a timeframe you allow, the more stuff would get swapped out for modern components until, Thesean Ship style, you'd end up with a design using mostly modern parts.
More realistically, engineers now probably wouldn't start from Apollo, but from more modern rocket programs, but the principle is the same: you can rush something out and have an exponentially increasing chance of mission failure, or you can go slower and steadier, and be progressively more certain of being able to do the mission successfully and safely.