We address the problem of evaluation of the client’s qualitative attributes made by the bank in lending credit. Till now, the banks have dodged to face this part of the lending credit. There are several reasons for this. One is the impossibility of using a statistical approach; the variables are linguistic attributes, not numbers. Another one, which we think really serious, is the difficulty of fixing which qualitative attributes are important. Every bank uses a personal contact with the client, the experts have not a unique behaviour. Here we present a sketch of our work, performed with an Italian bank, in which a fuzzy approach is used. In particular, we have used two different methods: a fuzzy expert system and a fuzzy cluster method.