.svg)
.svg)
Bots are the new black! The entire tech industry seems to be buzzing with “bot” fever. Me and my co-founders often see a “bot” company and discuss its business model. Chirag Jog has always been enthusiastic about the bot wave while I have been mostly pessimistic, especially about B2C bots. We should consider that there are many types of “bots” — chat bots, voice bots, AI assistants, robotic process automation(RPA) bots, conversational agents within apps or websites, etc.
Over the last year, we have been building some interesting chat and voice based bots which has given me some interesting insights. I hope to lay down my thoughts on bots in some detail and with some structure.
Bots are software programs which automate tasks that humans would otherwise do themselves. Bots are developed using machine learning software and are expected to aggregate data to make the interface more intelligent and intuitive. There have always been simple rule-based bots which provide a very specific service with low utility. In the last couple of years, we are seeing emergence of intelligent bots that can serve more complex use-cases.
Machine learning, NLP and AI technologies have matured enabling practical applications where bots can actually do intelligent work >75% of the times. Has general AI been solved? No. But is it good enough to do the simple things well and give hope for more complex things? Yes.
Secondly, there are billions of DAUs on Whatsapp & Facebook Messenger. There are tens of millions of users on enterprise messaging platforms like Slack, Skype & Microsoft Teams. Startups and enterprises want to use this distribution channel and will continue to experiment aggressively to find relevant use-cases. Millennials are very comfortable using the chat and voice interfaces for a broader variety of use-cases since they used chat services as soon as they came online. As millennials become a growing part of the workforce, the adoption of bots may increase.
Thirdly, software is becoming more prevalent and more complex. Data is exploding and making sense of this data is getting harder and requiring more skill. Companies are experimenting with bots to provide an “easy to consume” interface to casual users. So non-experts can use the bot interface while experts can use the mobile or web application for the complex workflows. This is mostly true for B2B & enterprise. A good example is how Slack has become the system of engagement for many companies (including at @velotiotech). We require all the software we use (Gitlab, Asana, Jira, Google Docs, Zoho, Marketo, Zendesk, etc.) to provide notifications into Slack. Over time, we expect to start querying the respective Slack bots for information. Only domain experts will log into the actual SaaS applications.
Consumer focused bots use popular messaging and social platforms like Facebook, Telegram, Kik, WeChat, etc. Some examples of consumer bots include weather, e-commerce, travel bookings, personal finance, fitness, news. These are mostly inspired by WeChat which owns the China market and is the default gateway to various internet services. These bots show up as “contacts” in these messenger platforms.
Strategically, the B2C bots are basically trying to get around the distribution monopoly of Apple & Google Android. As many studies have indicated, getting mobile users to install apps is getting extremely hard. Facebook, Skype, Telegram hope to become the system of engagement and distribution for various apps thereby becoming an alternate “App Store” or “Bot Store”.
I believe that SMS is a great channel for basic chatbot functionality. Chatbots with SMS interface can be used by all age groups and in remote parts of the world where data infrastructure is lacking. I do expect to see some interesting companies use SMS chatbots to build new business models. Also mobile bots that sniff or integrate with as many of your mobile apps to provide cross-platform and cross-app “intelligence” will succeed — Google Now is a good example.
An often cited example is the DoNotPay chatbot which helps people contest parking tickets in the UK. In my opinion, the novelty is in the service and it’s efficiency and not in the chatbot interface as such. Also, I have not met anyone who uses a B2C chatbot even on a weekly or monthly basis.
Enterprise bots are available through platforms and interfaces like Slack, Skype, Microsoft Teams, website chat windows, email assistants, etc. They are focused on collaboration, replacing/augmenting emails, information assistants, support, and speeding up decision-making/communications.
Most of the enterprise bots solve niche and specific problems. This is a great advantage considering the current state of AI/ML technologies. Many of these enterprise bot companies are also able to augment their intelligence with human agents thereby providing better experiences to users.
Some of the interesting bots and services in the enterprise space include:
Amazon had a surprise hit in the consumer electronics space with their Amazon Echo device which is a voice-based assistant. Google recently releases their own voice enabled apps to complete with Echo/Alexa. Voice assistants provide weather, music, searches, e-commerce ordering via NLP voice interface. Apple’s Siri should have been leading this market but as usual Apple is following rather leading the market.
Voice bots have one great advantage- with miniaturization of devices (Apple Watch, Earpods, smaller wearables), the only practical interface is voice. The other option is pairing the device with your mobile phone — which is not a smooth and intuitive process. Echo is already a great device for listening to music with its Spotify integration — just this feature is enough of a reason to buy it for most families.
Bots are useful and here to stay. I am not sure about the form or the distribution channel through which bots will become prevalent. In my opinion, bots are an additional interface to intelligence and application workflows. They are not disrupting any process or industry. Consumers will not shop more due to chat or voice interface bots, employees will not collaborate as desired due to bots, information discovery within your company will not improve due to bots. Actually, existing software and SaaS services are getting more intelligent, predictive and prescriptive. So this move towards “intelligent interfaces” is the real disruption.
So my concluding predictions:
If you’d like to chat about anything related to this article, what tools we use to build bots, or anything else, get in touch.
Did you like the blog? If yes, we're sure you'll also like to work with the people who write them - our best-in-class engineering team.
We're looking for talented developers who are passionate about new emerging technologies. If that's you, get in touch with us.
Bots are the new black! The entire tech industry seems to be buzzing with “bot” fever. Me and my co-founders often see a “bot” company and discuss its business model. Chirag Jog has always been enthusiastic about the bot wave while I have been mostly pessimistic, especially about B2C bots. We should consider that there are many types of “bots” — chat bots, voice bots, AI assistants, robotic process automation(RPA) bots, conversational agents within apps or websites, etc.
Over the last year, we have been building some interesting chat and voice based bots which has given me some interesting insights. I hope to lay down my thoughts on bots in some detail and with some structure.
Bots are software programs which automate tasks that humans would otherwise do themselves. Bots are developed using machine learning software and are expected to aggregate data to make the interface more intelligent and intuitive. There have always been simple rule-based bots which provide a very specific service with low utility. In the last couple of years, we are seeing emergence of intelligent bots that can serve more complex use-cases.
Machine learning, NLP and AI technologies have matured enabling practical applications where bots can actually do intelligent work >75% of the times. Has general AI been solved? No. But is it good enough to do the simple things well and give hope for more complex things? Yes.
Secondly, there are billions of DAUs on Whatsapp & Facebook Messenger. There are tens of millions of users on enterprise messaging platforms like Slack, Skype & Microsoft Teams. Startups and enterprises want to use this distribution channel and will continue to experiment aggressively to find relevant use-cases. Millennials are very comfortable using the chat and voice interfaces for a broader variety of use-cases since they used chat services as soon as they came online. As millennials become a growing part of the workforce, the adoption of bots may increase.
Thirdly, software is becoming more prevalent and more complex. Data is exploding and making sense of this data is getting harder and requiring more skill. Companies are experimenting with bots to provide an “easy to consume” interface to casual users. So non-experts can use the bot interface while experts can use the mobile or web application for the complex workflows. This is mostly true for B2B & enterprise. A good example is how Slack has become the system of engagement for many companies (including at @velotiotech). We require all the software we use (Gitlab, Asana, Jira, Google Docs, Zoho, Marketo, Zendesk, etc.) to provide notifications into Slack. Over time, we expect to start querying the respective Slack bots for information. Only domain experts will log into the actual SaaS applications.
Consumer focused bots use popular messaging and social platforms like Facebook, Telegram, Kik, WeChat, etc. Some examples of consumer bots include weather, e-commerce, travel bookings, personal finance, fitness, news. These are mostly inspired by WeChat which owns the China market and is the default gateway to various internet services. These bots show up as “contacts” in these messenger platforms.
Strategically, the B2C bots are basically trying to get around the distribution monopoly of Apple & Google Android. As many studies have indicated, getting mobile users to install apps is getting extremely hard. Facebook, Skype, Telegram hope to become the system of engagement and distribution for various apps thereby becoming an alternate “App Store” or “Bot Store”.
I believe that SMS is a great channel for basic chatbot functionality. Chatbots with SMS interface can be used by all age groups and in remote parts of the world where data infrastructure is lacking. I do expect to see some interesting companies use SMS chatbots to build new business models. Also mobile bots that sniff or integrate with as many of your mobile apps to provide cross-platform and cross-app “intelligence” will succeed — Google Now is a good example.
An often cited example is the DoNotPay chatbot which helps people contest parking tickets in the UK. In my opinion, the novelty is in the service and it’s efficiency and not in the chatbot interface as such. Also, I have not met anyone who uses a B2C chatbot even on a weekly or monthly basis.
Enterprise bots are available through platforms and interfaces like Slack, Skype, Microsoft Teams, website chat windows, email assistants, etc. They are focused on collaboration, replacing/augmenting emails, information assistants, support, and speeding up decision-making/communications.
Most of the enterprise bots solve niche and specific problems. This is a great advantage considering the current state of AI/ML technologies. Many of these enterprise bot companies are also able to augment their intelligence with human agents thereby providing better experiences to users.
Some of the interesting bots and services in the enterprise space include:
Amazon had a surprise hit in the consumer electronics space with their Amazon Echo device which is a voice-based assistant. Google recently releases their own voice enabled apps to complete with Echo/Alexa. Voice assistants provide weather, music, searches, e-commerce ordering via NLP voice interface. Apple’s Siri should have been leading this market but as usual Apple is following rather leading the market.
Voice bots have one great advantage- with miniaturization of devices (Apple Watch, Earpods, smaller wearables), the only practical interface is voice. The other option is pairing the device with your mobile phone — which is not a smooth and intuitive process. Echo is already a great device for listening to music with its Spotify integration — just this feature is enough of a reason to buy it for most families.
Bots are useful and here to stay. I am not sure about the form or the distribution channel through which bots will become prevalent. In my opinion, bots are an additional interface to intelligence and application workflows. They are not disrupting any process or industry. Consumers will not shop more due to chat or voice interface bots, employees will not collaborate as desired due to bots, information discovery within your company will not improve due to bots. Actually, existing software and SaaS services are getting more intelligent, predictive and prescriptive. So this move towards “intelligent interfaces” is the real disruption.
So my concluding predictions:
If you’d like to chat about anything related to this article, what tools we use to build bots, or anything else, get in touch.
.svg)
.svg)
