The era of all-sky space astrometry began with the Hipparcos mission in 1989 and provided the first very accurate catalogue of apparent magnitudes, positions, parallaxes and proper motions of 120 000 bright stars at the milliarcsec (or milliarcsec per year) accuracy level. Hipparcos has now been superseded by the results of the Gaia mission. The second Gaia data release contained astrometric data for almost 1.7 billion sources with tens of microarcsec (or microarcsec per year) accuracy in a vast volume of the Milky Way and future data releases will further improve on this. Gaia has just completed its nominal 5-year mission (July 2019), but is expected to continue in operations for an extended period of an additional 5 years through to mid 2024. Its final catalogue to be released ∼ 2027, will provide astrometry for ∼ 2 billion sources, with astrometric precisions reaching 10 microarcsec. Why is accurate astrometry so important? The answer is that it provides fundamental data which underpin much of modern observational astronomy as will be detailed in this White Paper. All-sky visible and Near-InfraRed (NIR) astrometry with a wavelength cutoff in the K-band is not just focused on a single or small number of key science cases. Instead, it is extremely broad, answering key science questions in nearly every branch of astronomy while also providing a dense and accurate visible-NIR reference frame needed for future astronomy facilities.