Virtual Agents, ChatBots, Invisible Apps and Conversation Commerce
Since drafting this blog post in Evernote a few months ago, the ChatBot hype has escalated. You will read about Virtual Agents, ChatBots, Invisible Apps, Conversation Commerce, Conversational UIs, Ambient Computing, Bot Stores, Tay and more. Checkout “The Complete Beginner’s Guide to ChatBots” on Medium for tons of examples.
Messaging apps like Kik, WeChat, WhatsApp, Facebook M, Slack and Telegram are integrating ChatBots into their experiences and building Developer tools. Experiences for Shopping, Booking Travel, Dating, Marketplaces and Healthcare are being built by Developers, Startups and Enterprises.
Over the past year since first learning about “Invisible Apps” and ChatBots I’ve been following the thinking of Chris Messina. Chris is a thought-leader on this subject. I highly recommend checking out these three writings/podcasts/lists from him:
- 2016 will be the year of Conversation Commerce
- Intercom podcast with Chris Messina
- Chris Messina’s Product Hunt Collection
After you checkout Chris’s work and get a sense of what’s happening, read on…
Designing for AI, Bots and Invisible Apps
One of the aspects of this I find fascinating is the psychology behind bot interactions. My wife, 7 yr old and 5 yr old and me all say “Thanks Alexa” after “she” does something like sets a timer or plays a song. I think most of us have had interactions in which we’re not sure if we’re talking with a bot or a human.
These articles and podcasts discuss the design, psychology and product management approaches to designing for this new paradigm.
- Dennis Mortensen, Founder of x.ai
- Stelios Constantinides, UX Designer/Engineer at Truth Labs
- William Peng has a nice piece on thinking about Platform Bots
- Benedict Evans talks about Chat bots, conversation and AI as an interface
Dev Platforms for Bots
Twilio is totally awesome of course and has been around for a while. We also see startups like Chat Fuel, Layer and Button supporting the new bot type development and technologies like IFFT being leveraged.
Microsoft is also pushing hard in this area by providing a Developer Framework called BotBuilder. And Facebook’s Messenger Platform of course.
Why is this important?
On average, people spend the majority of their time in about 5 apps and interact with about 25 each month. Most experts are predicting a trend in which other interactions beyond apps gradually change the app paradigm that has dominated the past 5+ years. Interacting with notifications, SMS, messenger clients, Siri, Google Now, Cortana, Alexa are all examples of this. Of the experts that predict the “death of apps”, there’s the same amount that predict “ChatBots are hype and apps as we know them are here to stay”.
The “Second Smartphone Revolution” predicts people will bank, shop and interact with their doctor as core mobile experiences. With the mobile phone as the primary device for these important day-to-day interactions, you can imagine how conversational commerce and ChatBot type experiences could play a big role.
Even More Reading
Here’s a few more podcasts and articles I recommend on this topic.
- Contextual Runtimes by Fred Wilson
- Re/code podcast with John Borthwick
- Intercom podcast with Paul Abrams
For Me Personally
Some of my favorite user experiences now are coming from apps that interact mostly via SMS including Digit and TripIt.
I feel myself gradually getting very comfortable with alternative UIs like ChatBots and Alexa. I trust the information I’m submitting is being handled correctly, I enjoy interacting outside of a website, email or mobile app and I find myself experimenting with different types of things. A few times a day I have an interaction where I can see how it could and will be improved by a “Invisible App” type experience.
Who knows if ChatBots will be the next best thing since sliced bread, but it’s certainly a cool paradigm that I see as having an important role in our daily lives.
Getting Started With AlchemyAPI
AlchemyAPI is a collection of APIs that help you understand text and images.
Here are three links to checkout before moving on:
Get your free API key
Alright, now that you have your API key, you can continue on with these examples. Simply replace your API key and copy into a browser tab.
Working with Text:
Using the “combined” call, you can extract a large amount of meta data from any document or URL including entities, relations, concepts, sentiment, taxonomy and more.
// replace YOUR API KEY http://access.alchemyapi.com/calls/url/URLGetCombinedData?apikey=YOUR_API_KEY&outputMode=json&knowledgeGraph=1&extract=page-image,image-kw,feed,entity,keyword,title,author,taxonomy,concept,relation,pub-date,doc-sentiment&url=https://www.reddit.com/r/announcements/comments/3djjxw/lets_talk_content_ama/
To get “clean” text from a web page by removing ads and other unnecessary content:
// replace YOUR API KEY http://access.alchemyapi.com/calls/url/URLGetRawText?apikey=YOUR API KEY&url=http://www.cnn.com/2009/CRIME/01/13/missing.pilot/index.html
Working with Images:
To find objects and text within an image, combine these two AlchemyVision API calls:
// replace YOUR API KEY http://access.alchemyapi.com/calls/url/URLGetRankedImageSceneText?apikey=YOUR API KEY&outputMode=json&url=http://www.coloradoski.com/uploads/Telluride_Helitrax_brett_schreckengost_logos.jpg http://access.alchemyapi.com/calls/url/URLGetRankedImageKeywords?apikey=YOUR API KEY&outputMode=json&url=http://www.coloradoski.com/uploads/Telluride_Helitrax_brett_schreckengost_logos.jpg
To find demographics of the people within an image, use the Face call:
// replace YOUR API KEY http://access.alchemyapi.com/calls/url/URLGetRankedImageFaceTags?apikey=YOUR API KEY&outputMode=json&knowledgeGraph=1&url=https://instagram.com/p/34XsRcxCAR/
Querying the News:
To get news articles about IBM over the past 24 hours:
// replace YOUR API KEY http://access.alchemyapi.com/calls/data/GetNews?apikey=YOUR API KEY&outputMode=json&start=now-1d&end=now-0d&maxResults=10&q.enriched.url.enrichedTitle.entities.entity=|text=ibm,type=company|&label_format_string=enriched.url.url,enriched.url.title
and to find approximately how many articles were published by the WSJ over the past 30 days grouped by day:
// replace YOUR API KEY http://access.alchemyapi.com/calls/data/GetNews?apikey=YOUR API KEY&outputMode=json&start=now-30d&end=now&timeSlice=1d&q.enriched.url.url=wsj
About Language Support:
AlchemyAPI provides named entity extraction capabilities in 8 different languages: English, French, German, Italian, Portuguese, Russian, Spanish, and Swedish. View full Language Support