Post on 13-Apr-2017
Messengers, Bots and Personal Assistants
Konstantin Savenkov, CEO Intento
Data Science Week, Moscow,2016
Top options for any Data Scientist in 2016
1.
2. do something else
Data Science Week, Moscow,2016
Three trends to track
1. The rise of Chat Bots
2. Personal Assistants
3. API / Micro-services
IntelligentApps
More stuffaccessibleto 3rd partysoftware
Data Science Week, Moscow,2016
Trend I: The rise of chat bots• 2015: More people in messengers than in
social networks, more time spent in messengers on mobile than in web and social networks combined.
• Businesses: follow the audience, need tools
• Messengers: want to earn money without hurting the audience
• Chat bots: an attempt to turn a messenger into a smartphone / browser etc ?
Data Science Week, Moscow,2016
Trend I: The rise of chat bots
Data Science Week, Moscow,2016
2014 2015 2016
(apps, not bots)
NDA
(apps, not bots)
Some scepticism:a lot of money invested low to none revenues low user adoption
Are chatbots any good?NOT when they are trying to implement GUI browsing via chat dialog
Data Science Week, Moscow,2016
YES when they can operate based on a limited input, just like we humans
100 011011001
101001001001000 110010100001
10110
•GUI-over-Chat is no better than PPP-over-Voice
•e.g. replace existing services with chat interfaces: concierge, receptionist, waiter, psychoterapist, customer support, assistant etc
manual
Trend II: Personal Assistants• Another way to be as close to users as possible,
and to control his consumer behaviour
• The next best option after a hardware-locked app ecosystem, which may be obsolete soon enough
Data Science Week, Moscow,2016
2011 2015 20162012 2013 2014
Trend III: API / Micro-services
Service Provider
WWW
3rd partyappAPI
• API Management market CAGR 22% (Forrester)• Google just bought Apigee for $625M
Data Science Week, Moscow,2016
Web sit
es
Web ap
ps
Native
apps
Chat b
ots
Person
al
assis
tants
IoT
Interface apps Intelligent apps
App Evolution
Robots
(enable to do smth.) (enable to avoid doing smth.)
Data Science Week, Moscow,2016
Dumb app Intelligent app
Menu Waiter
Data Science Week, Moscow,2016
Intelligent App
Terminal
NLP/NLU
Reasoning
Service platform
Data Science Week, Moscow,2016
Data Science Chalenges:
• Understand natural language
• Infer context
• Take decisions
• Cooperate
Terminals• Chat and assistant terminals
• Connector services • Microsoft Bot Framework • API.ai • message.io
Data Science Week, Moscow,2016
NLP/NLU
Data Science Week, Moscow,2016
• IBM Watson Language Services
• Microsoft Cognitive Services LUIS
• Google Natural Language API
• WIT.ai
• API.ai
• angel.ai
Reasoning
• Alexa Skill Kit: set of sample queries -> provided API
• Viv: Automatic synthesis of the intent fulfilment pipeline
Service Platforms
• Manual: Google Now / Assistant
• Crowdsourced: Amazon Alexa Skill Kit, Apple Siri Kit, Viv + a lot in verticals (hotels, taxi etc)
• Automatic: Intento (coming soon)
Data Science Week, Moscow,2016
http://venturebeat.com/2016/08/11/introducing-the-bots-landscape-170-companies-4-billion-in-funding-thousands-of-bots/
Chat bots vs. Assistants
Data Science Week, Moscow,2016
Chatbot paradigm: dumb terminals clever services
Personal Assistant paradigm:
clever terminals dumb services
•
• •
••
• •
•
• •
Q&A
Konstantin SavenkovCEO Intento
<ks@inten.to>
Data Science Week, Moscow,2016