The goal of any cryptographic system is the exchange of information among the intended users without any leakage of information to others who may have unauthorized access to it. A common secret key could be created over a public channel accessible to any opponent. Neural networks can be used to generate common secret key. In case of neural cryptography, both the communicating networks receive an identical input vector, generate an output bit and are trained based on the output bit. The two networks and their weight vectors exhibit a novel phenomenon, where the networks synchronize to a state with identical time-dependent weights. The generated secret key over a public channel is used for encrypting and decrypting the information being sent on the channel. This secret key is distributed to the other vendor efficiently by using an agent based approach.