Case for Chatbots in IoT

The connected platform for things have been called many different names. One of the earliest one and mostly incorrect term was M2M or machine to machine. We want to connect coffee machines, vehicles, refrigerators, cameras and in the end connect it to humans. In all IoT applications things, data, connectivity are important but the human is the primary player. She is the final user whose needs are paramount. The aim of any IoT application is to make it easy for the user to manage the asset. We believe the chatbot presents a novel and effective interface for an user of an IoT product.

Who would use a Chatbot

There are several customers that would find the a fully conversational chatbot convenient, and the preferred mode of interaction. We believe the chatbot will have to multilingual, conversational and voice enabled to be truly popular.

Our earliest customers were tech savvy logistics startups. Their team talented team wanted to access data collected from their vehicles using APIs, MQTT and webhooks. Some of our customers wanted the web application or mobile application to monitor their vehicles.

One of our customers is one of the biggest insurance providers for financed commercial vehicles. About 85% of the commercial vehicles in India are bought by small business owners or individual entrepreneurs who own 1-20 vehicles. When onboarded them to our app we realized

1. Most of them could not read English

2. They were not technically savvy unlike our early adapters

3. Some of them did not have smart phones. They bought smartphones because they felt saw value in being able to track their vehicles

Since most of them were comfortable in Hindi, we released a multilingual version of our app that supported both Hindi and English. Most of them would not use our web interface, because they did not have PCs. The mobile app was a hit with them. Based on our observations, the users of a bot would be the non tech savvy customers, who is more comfortable on WhatsApp and Messenger.

Chatbot versus Mobile App

I briefly touched upon this subject previously. A mobile app has been the default mode for interaction on a mobile phone. A bot on an messaging app which the user is already comfortable with, will have less friction in adoption. There is often a resistance from a customer to download another app on to her device. The device she may be using if often low on memory and so they decide judiciously whether she wants another work related app on her phone. A chatbot on an existing Internet Messenger does not consume additional memory.

Notifications

In an IoT application, there often are issues that need urgent attention and so have to be escalated quickly. The popular methods of such escalations are

1. Email

2. SMS

3. App Notification

SMS comes at an incremental cost. Since we serve a very cost sensitive clientele we have always refused SMS based escalations. Email may not be dead, but is surely dying is favor of instant messaging services. App notifications are fine, as long as the user has the app on the phone. By using a chatbot on an incumbent instant messaging platform, she can be notified of issues immediately. On the Facebook Messenger bot, admins can get notified, when the following happen:

  1. Vehicle breaks a preset geofence
  2. Device gets tampered with
  3. Vehicle meets an accident
  4. Driver parks the vehicle in an insecure location
  5. Vehicle develops a problem and produces a DTC
  6. Driver drives the vehicle harshly

On receiving these escalations, an admin can do the following

1. Query its latest location

2. Remote lock-down the vehicle

We also use the Messenger internally within the dev-ops team to escalate any issues with out servers and applications. It has enabled our engineers to quickly respond to issues and resolve them before it becomes a problem for any of our customers.

You can read more about our Messenger Notification system and how we implement it in this blog.

Summary

The chatbot interface is going to be a game changer for IoT applications. The bots will have to become truly conversational to become popular. I have not come across a perfect chatbot yet, but I am continually amazed at what the speed they are evolving. They are quite sufficient for several applications already. They will also need to be multilingual and by piggybacking on incumbent instant messaging services, it can reduce the friction and become the primary means of interaction.

In case you want to explore how to use chatbots or our IoT application you can drop a comment and get in touch with us.