I cannot think of any specific encounter in which information or knowledge clearly improved your interaction with a human agent. Knowing less than the human agent keeps the interaction simple because knowing less than the agent can’t lead to dissatisfaction. I have found that in many cases where I have had access to knowledge about a product and information about company policy interaction with a human agent is useless. Many agents knew less than I did because they only dealt with the script given to them. Many times when I was knowledge agents ended up transferring me all around departments trying to find someone who knew what I knew and could help me.
The best example I recall is in my dealings with Time Warner Cable. Every time my Internet is down I call Time Warner Cable, let them know and the first thing they tell me every single time is that I can go online to their website for assistance. This is frustrating! No one wants to sit on the phone for one hour on hold trying to get their paid service back for someone to come on and tell you you should go online for help first. This is bad customer service. On top of this human agents at Time Warner Cable indefinitely can’t help you because they are always transferring the customer to another department.
Human interaction when it comes to customer service is vital from a reasoning perspective, humans you would think, could decipher problems and help the customer work through it but, the realty is CS agents really don’t help much outside of recording issues. Internet Automation is beginning to become more sophisticated in the U.S. allowing for self service, self education and self scheduling. Humans are phasing themselves out.