Detecting exoplanets in clusters of different ages is a powerful tool for understanding a number of open questions, such as how the occurrence rate of planets depends on stellar metallicity, on mass, or on stellar environment. We present the first results of our HARPS long-term radial velocity (RV) survey which aims to discover exoplanets around intermediate-mass (between ~ 2 and 6 Msun) evolved stars in open clusters. We selected 826 bona fide HARPS observations of 114 giants from an initial list of 29 open clusters and computed the half peak-to-peak variability of the HARPS RV measurements, namely DeltaRV/2, for each target, to search for the best planet-host candidates. We also performed time series analysis for a few targets with enough observations to search for orbital solutions. Although we attempted to rule out the presence of binaries on the basis of previous surveys, we detected 14 new binary candidates in our sample, most of them identified from a comparison between HARPS and CORAVEL data. We also suggest 11 new planet-host candidates based on a relation between the stellar surface gravity and DeltaRV/2. Ten of the candidates have less than 3 Msun, showing evidence of a low planet occurrence rate for massive stars. One of the planet-host candidates and one of the binary candidates show very clear RV periodic variations, allowing us to confirm the discovery of a new planet and to compute the orbital solution for the binary. The planet is IC 4651 9122b, with a minimum mass of msini = 6.3 MJ and a semi-major axis a = 2.0 AU. The binary companion is NGC 5822 201B, with a very low minimum mass of msini = 0.11 Msun and a semi-major axis a = 6.5 AU, which is comparable to the Jupiter distance to the Sun.