Conversations Squared #8 - Michelle Parayil

Let’s talk about proactive messaging! Michelle Parayil, Conversation Design Manager at Haptik, knows all about it. She’s learnt a great deal about proactive messaging with conversational AI, and shares her learnings and experiences. When’s the best time to do proactive messaging? How do we avoid a notification pile-up when sending chat messages proactively? Can a bot that’s too proactive create a bad user experience? It’s all in there, and more… A wee note about this episode: Michelle called while she was in the middle of a thunderstorm! That’s dedication for you. There’s a few booms and bangs in the background, but she’s still loud and clear. Hope you enjoy it. 00:00 Start 02:08 What's the most useful thing you use at work? 03:22 How do you select the right use case? 04:24 Should we aim for bots to mimic humans? 06:32 What is proactive messaging? 08:14 What are the challenges and benefits of proactive messaging? 11:30 Is there a danger of a notification pile-up? 13:30 Can proactive messaging be done 24/7? 15:26 Can an overly proactive bot create a terrible experience? 17:38 How does data influence the time when we message the user? 20:21 Are there different ways to interpret the user's explicit 'no' to proactive messaging? 23:05 How can a bot fallback if they proactively messaged at the wrong time? 24:31 When would you recommend to not use proactive messaging? 26:28 Does proactive messaging create a strong impression of the bot's persona? 27:21 In which channels do Haptik use proactive messaging? 30:59 What do you think is the future of proactive messaging? 33:13 Outro