The dual-channel UV spectrometer VenSpec-U onboard ESA’s next mission to Venus, EnVision, will study the upper layer of the atmosphere located above the clouds. It aims to characterise the chemical composition, with a focus on the sulphured gases (SO2 and SO) and the identification of the unknown UV absorber, and will also monitor the dynamical processes such as gravity waves and convection cells. In this article, we are interested in knowing how the instrumental design allows to comply with requirements arising from these scientific objectives. This study is based on the radiative transfer model (RTM) developed for SPICAV/Venus-Express data analysis, that is used to retrieve of atmospheric features from radiance factor spectra, which will be derived from VenSpec- U’s measurements of Venus’ radiance. We will then study the sensitivity of the model to various error sources: random errors or biases. For the first ones, we assess the impacts of the Signal-to-Noise Ratio on the uncertainties of the inverse RTM outputs, in order to check that the required uncertainties are achieved for the main science goals. Limits in terms of SNR can also be defined in order to ensure the compliance with the specifications. We then present the approach implemented for the characterisation of systematic errors: the Effective Spectral Radiometric Accuracy (ESRA) requirement is used to estimate the impact of a bias on the accuracy of the retrieved science products based on the spectral characteristics of the biased spectra. After identifying some bias sources that could occur in VenSpec-U’ case, combinations are considered in order to study potential compensations and estimate allowable envelopes of residual error levels for each kind.