Solution NMR of paramagnetic metal complexes contains a lot of useful information on their geometry, dynamics, electronic structure and magnetic properties. However, it is often impossi- ble to retrieve this information due to difficulties in spectra assignment that complicate further analysis. In our work, we show how computational chemistry can help to analyse pNMR spec- tra and extract full magnetic susceptibility tensor from standard 1H, 13C NMR. Our case study of radical-bridged bimetallic complexes of Ni(II) and Co(II) demonstrates that variable tem- perature solution NMR is able to constrain more spin-Hamiltonian parameters than commonly employed powder SQUID measurements.