What you want to do is the same reddit did: sock puppets; but ethically sourced. Write a LLM routine that looks at a neutral source and starts using your product.
People generally dont know what to do when they can do anything.
Also, once you get active users, your get active trolls, so you'll need some kind of active or passive denial system. I think the most successful system is a lifetime $ subscription cost and a thorough banning of trolls. Either you make a lot of money on troll costs or you end up with a nice harmony.
That being said, we will have a feature soon where you can compose the entire message before sending it (the 'traditional' way).