Skin sensitisation is an important toxicological endpoint - both for the purposes of protecting workers at their workplace, and to ensure that the general population doesn't get sensitised. Testing on animal models has been performed for decades. However, several years ago, animal testing for cosmetic ingredients has been forbidden in Europe and this trend is expanding to other countries. In the '90 the development of non- animal method for skin sensitisation has started and several methods have achieved regulatory acceptance. Because the single test methods are covering only one key event in the AOP it is thought that more than one assay would be necessary to sufficiently cover the AOP and to replace the animal models. Here, I address the questions regarding the required building blocks for a skin sensitisation risk assessment based on non-animal methods and approaches. Due to the variety of the needs and the complexity of risk assessment in terms of the data that can feed the models of risk assessment for skin allergy, these building blocks should to be both flexible and clear to provide the right information. For this reason, I present a classical framework based on ICCR principles and the Chemical safety assessment workflow based on exposure considerations and non-animal methods from the OECD. According to the key elements of the framework, the question on risk assessment must be clear, be human relevant, exposure led, hypothesis driven and designed to prevent harm. A tiered iterative approach should be applied, starting with a thorough review of all of the available existing information. The data from non-animal test methods can be used in a defined approach (DA) and the chosen DA might be dependent on the available information and the question on risk assessment. The framework should also include the elements from in silico data, as well as the bioavailability data, which are key points in the initial screening and the refinement of exposure, respectively. In conclusion the risk assessment can be conducted successfully following the framework with “traditional” animal data and NAMs.