Funny (or frightening) piece of news on Monday, guys!

During Super Bowl game, Google’s ad for the Google Home aired on millions of TVs: loving families at home meeting, hugging, and being welcomed by the Google Assistant. Someone says “Okay, Google,” and that’s one has triggered a whole bunch of Google Assistants all over the United States. This isn’t the first time television has set off people’s home assistants. A month ago, a TV broadcast accidentally triggered a whole bunch of Amazon Echos.

At some point, some enterprising TV writer or ad jerk is gonna plant an “Okay, Google” on TV with intent and force everyone to listen to the music he wants or to buy something on Amazon. This is a massive troll waiting to happen.

https://goo.gl/wiiZs6

When we’ve been working on our “bots for celebrities” product called Virtz, our users were in love with it, the average session was quite long (7 minutes), but we faced one big issue - users were trying it once or twice, and then forgetting about it. We did try to implement different retention features like ability to subscribe to notification about celebrity new Instagram post or Youtube video, but people were furious when we sent them messages and blocked our bot. So, we definitely made a huge mistake in conversation design and fucked up our retention rate.

The article below is a similar story (but with the happy-end) from the team behind Meekan, a Slack bot for scheduling. There you can find out their steps to get from 0.5% to 60% retention rate.

https://goo.gl/ZEiaaq

I think that one of the best and simplest bot cases right now is gathering user attitude and thoughts about some news/events. And Washington Post media agree with me on that. They’ve build a simple Facebook Messenger bot called Feels to track people’s feelings about the USA President election. Bot messages you every day and asks a simple question: “How is the election making you feel today?”. You need to respond with one emoji from the predefined set - (😄 🙂 😐 😢 😠). And that’s it. Sounds like useless/boring stuff? A full 85% of users stayed subscribed throughout the entire experiment, aha.

In the early days users reported feeling mostly 😐, for a wide variety of different reasons. But emotions change around some significant points in the race.

This gif shows how reactions to the election changed on key dates

Article about the experiment

PayPal today launched its first bot for users of team messaging app Slack. The bot lets people send peer-to-peer (P2P) payments with the shortcode /paypal and a user’s Slack handle. Basically you can say in a chat: /Paypal send @lauren $6 for Tuesday’s Latte. Confirm transaction via tapping button on the webview and that’s it.

This is PayPal’s first bot but not its first venture into the emerging chatbot ecosystem. Last November, PayPal joined Square Cash and Venmo to allow P2P payments with Siri. Weeks earlier, PayPal and Stripe joined a Facebook Messenger beta program to accept payments through the platform.

Demo video
Announcement

I really like Intercom and here is an article from them about the future of chatbots with explanations why voice will be the dominant chatbot interface.

2016 showed us that texting chatbots is hard and takes time and effort. That’s why a lot of people right now tend to think that voice interface is a solution. But I think they are wrong. Good conversational design is a solution. As for me, I don’t see the voice to be the dominant chatbot interface, I think there’ll be more of a combination. While it’s super convenient to order a pizza with voice commands while driving a car, there are a plenty of situations when you don’t want to say something aloud.

I’ve shared a poll below to ask your opinion about this point, thanks for voting!

https://goo.gl/50cnEp

Will the voice be the dominant chatbot interface in 5 years?

Few days ago the Russian Orthodox Church released a new social network build on top of Telegram.

In this dedicated social network users can join the official channels of over 16 000 churches as well as connect to people within the same church. They can receive news, Bible readings and organise volunteer help.

Within the social network users can also use different chatbots available on Telegram. For example a chatbot which searches for a nearest pharmacy ect.

You can try the Telegram Social Network yourself here (available in 8 languages)

I think that Paypal integration to receive donations from volunteers will look interesting here:

/Paypal send @priest_michael 200 rubles for Tuesday’s candle

America new President really likes to use Twitter. He uses it to express strong opinions about all sorts of things. Occasionally, those opinions are targeted at publicly traded companies. After Trump negative tweets about some company their stock price abruptly drops.

So, one guy decided to write a bot that constantly monitors Trump’s feed and analyzes each tweet for mentions of publicly traded companies. So, after that you can immediately trade the affected stocks based on the detected sentiment: buy if positive and short if negative. Sounds like a big piece of fun, right?

Article about the experiment: https://goo.gl/LDa6a8

Also bot posts tweets out a summary each time it springs into action here on Twitter: https://goo.gl/J8ol7H
And even more - the bot is completely open source and available on Github: https://goo.gl/8kze2U

If you’ve ever been building a bot, you must have thought about ways to generate revenue from it. Here is a great article from Amir Shevat @ Slack about different revenue models for chatbots.

Actually, there are a lot of them:

1. Subnoscription model (Statsbot)
2. Ad serving (works not for all platforms, Slack doesn’t allow)
3. In-content ads (instant articles from Facebook)
4. Data analytics and market research (Swelly has a loooot of data about user choices)
5. Selling goods and services (Amazon Alexa to order products for kitchen by “Alexa, add to shopping list”)
6. Referral fees (different shopping assistants)
7. Brand promotion (H&M, Zara, BMW)
8. In-bot virtual goods (virtual coins in games)

And many more in the article below.

I hope it will help you choose the monetization model.

https://goo.gl/GDvKWl

Chris Messina (inventor of the hashtag, Lead Developer Experience at Uber) and Amir Shevat (Head of Developer Relations at Slack) are the two the most powerful influencers in the space of bots I suppose. And it’s always interesting to hear their thoughts and listen to their discussions about future of bots.

And if you’re tired of reading, here is a recent podcast with Chris Messina where they discuss MessinaBot and personal bots, "bot skeuomorphism", the role of Product Hunt in the bot ecosystem, and a bunch of other topics.

I’ve also attached the file with the podcast in the message below to listen directly in Telegram.

Happy listening :)

https://goo.gl/54SPXU

It’s Sunday evening here, so it’s time for #botoftheweek nomination.

This week I’ve found an interesting bot that can tell you a name of a plant. Just take a photo of any plant and send it to the bot, it will identify the plant for you in just one second. It is backed by a customized deep learning algorithm with tons of data and tuning for plants. And I can say it’s pretty accurate.

Screenshot: https://goo.gl/vyhnRM

Bot is here to check: https://goo.gl/uAhKVx

Hey folks, I think that one way connection with you is too boring, because I don’t get feedback from you, so I decided to create a public chat where we can discuss a huge amount of bot stuff, like messaging apps, conversational interfaces, AI, ML and much more! Chat language is English, so feel free to join and share it with your friends if they are interested in bots!

https://news.1rj.ru/str/joinchat/AAAAAEIJ70FNIceDnr0WIg - JOIN HERE
https://news.1rj.ru/str/joinchat/AAAAAEIJ70FNIceDnr0WIg - JOIN HERE
https://news.1rj.ru/str/joinchat/AAAAAEIJ70FNIceDnr0WIg - JOIN HERE

Chat cover photo

While working on our last bot, we were kinda surprised that users don’t always click buttons when they see buttons, don’t always respond in text when they see a question, but they tend to interact more with visual content like stickers, images, gifs etc.

Here is an interesting from Dashbot, they’ve processed almost 700 million messages through their platform and analyzed them to find a percentage of text, images, stickers and other types of messages. They’ve also processed all the images through the Amazon’s Rekognition API to find out what was pictured on them. Very interesting statistics to look through.

https://goo.gl/A4Vc32

News from the messengers side: Amazon unveils Chime, looks to reinvent the conference call with new Skype and GoToMeeting competitor. Pricing starts at $2.50 per user per month on the lowest end, with a higher tier plan of $15 per user which includes video and screen-sharing among other features. There is a basic option available for free but that is limited to video calls and chat rooms between two users.

Maxim keeps fingers crossed while dreaming about death of Skype 🙏🏻

https://goo.gl/GFGGwS

When designing a product for the web or mobile world, we usually operate within certain types of interactions and user interface elements — text fields, forms, buttons, checkboxes, or switches, for example. In the chatbot world, the interface elements are different.

I found a very interesting article, especially for bot novices, about all Facebook chatbot interactions. It has a very comprehensive and huge pdf with a visual summary of all the interaction elements that are now available for the Facebook Messenger bot. It’ll save you at least a couple of hours.

Article: https://goo.gl/G1MYQ2
Link to the pdf directly: https://goo.gl/GlngeY

Long read about AI today (it's not new, but worth sharing): Apple gave Steven Levy an inside look at some of its deep learning and artificial intelligence work. Apple is said to have many people working on AI across various company initiatives. Of course much of the work is highly secretive, but some of it is already reflected in current iOS features—and not just Siri.

If you're not available to read it today, just add to the Pocket to come back to it later at weekend ;)

https://goo.gl/75agTq

Have you even been dreaming about your chatbot getting featured on Facebook? It could be a total disaster if you’re not prepared!

Here is an article from team behind Nerdify bot about their experience in being featured which costed them a couple of sleepless nights :)

https://goo.gl/uJgNcR

Have your bot ever been featured? Tell us here.

It’s always been obvious for most of people that robots will take over different routine tasks of our daily lives, freeing us to work on more creative stuff. But, surprisingly, some of the easiest tasks to automate turn out to be creative ones (art, science/research, legal, policing/security) and many of the monotonous jobs we’d really like to automate are proving very tricky (cleaning, clothing, farming). Have you ever been thinking about this?

https://goo.gl/a6Tltr

Do you remember my post about decision trees vs web views? But what if there’s a much better way?

Here is a concept of building conversational experience called Random Access Navigation. It’s main idea to give people the ability to navigate without a defined path, while also being able to change their mind at any time. R.A.N. offers more than one solution for the most common bot criticism: fixed decision trees (everybody hates them).

Very interesting concept for bot builders.

Article: https://goo.gl/J7aRyp

Decision trees vs web views