Making a physical product that may have to be tailored to varying specific requirements could be difficult, meric above mentioned technology / skill transfer where an assumed solution may not have an optimal effect, or be even be disastrous - I'm sure they're honing an experience set and training and maintenance manual.
Sewage works may be a tangential example. Many sewage treatment plants basically work by sifting heavy material, then using aerobic bacteria to break down dirty water, and anaerobic bacteria to break down the aerobic bacteria (resulting in manure slosh) [a heavily simplified version]. But each plant has varying PH and mineral levels, meaning optimal bacteria are often designed for specific plant locations, and the specialist knowledge required can be very deep.
I see nothing wrong with going it slow, and documenting carefully in a very hands-on way, any hiccup and both 'known unknowns' and 'unknown unknowns' that may be overlooked in a mass roll-out.
[Edit: typo]