Which is (I assume) a true statement, but does not actually support your blanket statement. "For design purposes" really means "if you want to avoid calculating actual figures, then this will be near enough for common situations". U (or R) values for soil varies significantly, but it DOES have thermal resistance, and it is not "negligible".... for design purposes, you can consider the oversite concrete to be a constant 5 degrees.
And as I said, part of the reason it will appear to have exceedingly poor insulating qualities is the thermal mass that will suck heat away until equilibrium is reached. Put another way, if you did rely on a couple of meters of soil as insulation - you'd find yourself putting lots of heat in initially and might conclude that the soil has no insulating quality, but that heat isn't "gone", that thermal mass is now holding much of it.
And "I've designed and installed ..." is not automatically a qualification to be taken at face value*. If you've always assumed that soil is not an insulator, and you've always assumed that the oversite is a constant 5˚C, then you have probably never even considered the design of an earth insulated system. And if you've never looked at the numbers, then your experience is not relevant to the specific case in point.
Not that I plan on relying on soil for thermal insulation - it won't be necessary.
* Apologies, this really isn't aimed at yourself, but I've met a few "heating engineers" who I wouldn't trust to plug in a fan heater. And often the reason for "disagreement" is them relying on stuff they take for granted - like "it's illegal for a homeowner to touch anything to do with gas" (it isn't) and "the boiler must be connected with a fused spur, it's in the instructions" (I made them put the socket back when doing mum's boiler, and it wasn't). And my favourite was the guy who was 101% adamant that Grundfoss, DAB, Wilo (and I assume others) don't make these modern modulating pumps.