Anything that comes out of marketing carries likelihood of inflated specs if not downright lies. You don't have to go any further than consumer grade shop vac HP specs to find a criminally egregious example.
Machine tools are more responsibly rated but you still have to have you BS detector in good working order when making a purchase decision.
To me, "positioning accuracy" is the quantfiable result of an axis move as measured by suitable instrumentation. The move is within he tolerance band or it isn't so results are easy to evaluate.
"Machining accuracy" can be a cloudier concept because the "positioning accuracy" is overlaid by tool and part deflection, thermal expansion, etc each attribute in he equation carrying its own load of uncertainty. However, if a prospective machine could machine a series of consistent parts I'd be inclined to accept its value in a production setting provided it competed favorably with comparable machines is its niche.
I'd be inclined to go easy on the employment of cuctter comps etc in demonstrations focusing more on part to part consistency, ease of programming, tool changes, and maintenance plus a ready supply of repair parts, quick field service, and customer support - all essentials in post-purchase customer satisfaction. These after-market customer support features often make or break the success of integrating a new and complex machine tool in a working production shop.
I recall a stout looking machine tool manufactured in the late '80's by a fading giant of a US manufacturers. It featured many novel seeming innovations but they were poorly executed the tool changer was particularly troublesome in that it mis-presented tools, dropped tooling, the post processor was glitchy, and important stuff plain broke.
The company failed soon after leaving Uncle Sam with a white elephant the successor company was reluctant to support. The specs on this machine were excellent and were met at factory acceptance and one installation run-off. Trouble is, it was unreliable in many important particulars. A good idea poorly implemented in tragic ways. A $180,000 pity.