The presentation explores the current state of investing in AI, including its industrial split, and provides a detailed outlook on AI applications in Healthcare, Transportation and Industrial sectors.
2. Table of content
Summary and overview of data collection
Presenter’s bio
Definition of AI
Overview of technology progress
Investment trends
M&A activity highlights
Precedent investments by industry
Implementation – case studies
Conclusion: how to cope with AI disruption
3. Summary and overview of data collection
Summary
Defining AI as a set of technologies allows to describe it better. Core technologies, like computer vision and
machine learning developed rapidly in the recent years.
Investmentsin AI also grew, from 0.9% of all venture capital investmentsin 2012 to 3.1% in H1 2016.
Cross-industrial business applications, as well as specific applications in Healthcare,Manufacturingand
Industrial Services and Transportation/Logistics represent 57% of $1.6B invested in AI in H1 2016.
Successful implementations of AI tech are already available in industries like Telecommunications, Banking
and Manufacturing.
Data collection
Data on investment in AI was collected and analyzed by Flint Capital from diverse sources including
Pitchbook and Crunchbase.
251 company in the US, the UK and Canada were identifiedvia key-word search and manually checked. We
included only those companies which received venture capital financing and were privately held at the time
of our research.
4. Peter Zhegin
A venture capital professionaland an AI enthusiast
Associate at Flint Capital, an internationalventure capital fund with exposure to cognitive technologies.
Investmentsinclude:
• CyberX – machine learningfor Industrial IoT,
• Findo – an NLP-powered smart search engine,
• Epistema – collaborative knowledge analytics platform,
• AudioBurst – transcribes audio and understands the meaning of spoken words in real-time,
and others…
Co-lead at Russia.AI – a non profit initiative aiming to support Russian speaking AI entrepreneurs.
Previously: Ozon.ru, ABRT Venture Fund.
MSc in Management,Leeds UniversityBusiness School,
MA in History, Moscow City Pedagogical University.
5. Artificial Intelligence (AI) may be considered as a set of several technologies
Narrative definitions are too wide and they hardly describe what AI actually is
Cognitive technologies comprising AI (2)
36%
(1) Nils J. Nilsson, The Quest for Artificial Intelligence: A History of Ideas and Achievements (Cambridge, UK: Cambridge University Press, 2010).
Cited from: One Hundred Year Study on Artificial Intelligence (AI100),” Stanford University, accessed August 1, 2016.
(2) Deloitte University Pres, http://www.theatlantic.com/sponsored/deloitte-shifts/demystifying-artificial-intelligence/257/
For example: ’Artificial intelligence is that activity devoted to making machines intelligent, and intelligence
is that quality that enables an entity to function appropriately and with foresight in its environment’ (1).
AI as a set of technologies
6. 0
100
200
300
400
500
2015 Amazon’s Picking
Challenge champion
2016 Amazon’s Picking
Challenge champion
Human worker
Progress in technology is stunning
Robots are quickly mastering difficult tasks
# of items picked per hour
by a robot vs. a human worker (3)
(3) http://futurism.com/deep-learning-ai-leads-robot-to-victory-in-amazons-picking-challenge/
(4) http://www.economist.com/news/special-report/21700756-artificial-intelligence-boom-based-old-idea-modern-twist-not?frsc=dg%7Cd
The gap between a human
and a robot is still wide
3x progress
Error rate on ImageNet
visual recognition challenge, % (4)
0%
5%
10%
15%
20%
25%
30%
2011 2012 2013 2014 2015
Goes beyond
human level
7. Investments in AI have been growing rapidly in the last five years
AI investments grew from 0.9% of world’s venture capital in 2012 to 3.1% in H1 2016
Venture capital investments in AI and other sectors, FY 2012 - H1 2016, Worldwide,$B and % (5)
(5) AI excludes accelerator and incubator deals https://www.cbinsights.com/research-venture-capital-Q2-2016 , https://www.cbinsights.com/blog/artificial-intelligence-funding-trends/,
https://www.cbinsights.com/blog/artificial-intelligence-funding-trends-q216/
44.8
50.3
89.1
128.5
52.2
0.4
0.8
2.2
2.4
1.69
0.9%
1.5%
2.4%
1.8%
3.1%
0.0%
0.5%
1.0%
1.5%
2.0%
2.5%
3.0%
3.5%
0
10
20
30
40
50
60
70
80
90
100
110
120
130
140
FY 2012 FY 2013 FY 2014 FY 2015 H1 2016
VentureinvestmentsinAIas%oftotal
Ventureinvestments$B
Other venture investments AI venture investments AI venture investments as % of total
AI – 3%
of VC
invested
in tech =
$1.69B
Annual investment
in AI ~$2.4B
8. 24 AI tech companies were acquired or went public for the total disclosed value of $1.2B
Selected AI tech companies acquired in H1 2016
Corporations are interested in getting AI tech in H1 2016
Acquirers of AI tech
are very diverse in
terms of industries
and
business models
9. 531
263
189
170
107
102
99
61
33
24 24 16 15
Cross-industrial Consumer products/Services
Healthcare Infrastructure
Transportation/Logistics Manufacturing/Industrial services
Retail/Commerce Finance/Insurance/Legal
Agriculture Aerospace
Security/Defense Education
Construction/Maintenance/Utilities
Cross-industrial business functions like HR or Marketing attract AI investments
Healthcare, transportation,and manufacturing are approached by AI investors as well
Venture capital investments in AI by industry, H1 2016, US,UK, and Canada, $M (6)
(6) Data from Flint Capital
Applications of AI in
Marketing, HR, Business
Intelligence, Sales, and
Administrative functions
AI in manufacturing
processes and analytical
applications in
manufacturing,
warehousing
Self-driving cars for
consumer and business
purposes, route-planning
software
Medical imaging,
surgery robotics,
analytical software
Core technologies e.g.
computer vision
or NLP
Consumer applications:
toys & games, photo
editing, household
robotics
Cross-industrial
Consumer
Healthcare
Infrastructure
Transportation
Manufacturing
Every industry uses
AI in its own way
10. Data – the key component of healthcare and AI’s target
95% of investments in AI that helps to collect, analyze and predict
AI in Healthcare, sub-segments & examples, H1 2016, US, UK, and Canada, $M (7)
97
50
33
8 1
Data digitalisation/Interpretation
Analycal processes augmentation
Data-related processes automation
Manual+Cognitive processes augmentation
Interaction/Communication automation
Exoatlet*
Visiongate
Restoration robotics
(7) Data from Flint Capital
* When here and further a company is marked by ‘*’ – the company was not included in the sample pf H1 2016 investments and is used as an example
11. 103
4
Manual processes automation Analycal processes augmentation
Transportation – self-driving cars is the key topic
>95% of investments in Transportation segment were allocated to self-driving cars
AI in Transportation, sub-segments & examples, H1 2016, US, UK, and Canada, $M (8)
NAMI-Yandex-KamAz*
Clearmetal
Zoox
(8) Data from Flint Capital
12. 57
43
3
Analycal processes augmentation
Manual processes automation
Data digitalisation/Interpretation
Manufacturing/Industrial services are balancing between analytics and robotizing
AI is applied almost equally to analytical and operational elements of production process
AI in Manufacturing/Industrial services, sub-segments & examples, H1 2016, US, UK, and Canada, $M (9)
Seegrid
(9) Data from Flint Capital
RoboCV*
Senseye
CyberX*
14. Telco's, banking, steel production, consumer services – AI seems to be industry agnostic
Case-studies of AI tech implementations
(10) http://eprints.lse.ac.uk/64516/1/OUWRPS_15_02_published.pdf ,
(11) SDBA Group presentation, 2016
(12) https://yandexdatafactory.com/case-studies/ydfs-recommender-system-to-decrease-steelmaking-costs-at-magnitogorsk-iron-and-steel-works/
(13) Findo’s presentation, 2016
Telefonica Retail bank Magnitogorsk Iron & Steel Works
By Blue Prism (UK) By SBDA Group (RU) By Yandex Data Factory (RU)
Telefónica O2 automated 15 core
processes including SIM swaps,
credit checks, and others,
representing about 35 percent of
all back office.
FTEs had been reduced on the
automated processes by a few
hundred. UK-based people were
redeployed to other service areas
and the business continued to
grow (10).
ML enabled the bank to send to cli
ents very targeted messages releva
nt to their real life events.
As a result, usage of some services
increased e.g. from 1% to 6% for
paying parking fines
online and from 41% to 74% for
mobile top-up services (11).
Yandex Data Factory created a rec
ommender system, integrated into
MMK’s software, that helps to
reduce ferroalloy use by
an average of 5%.
This equates to annual savings of
more than $4m in production costs
(12).
Several industries have already started harvesting fruit of AI implementations
Consumers
By Findo (RU/US)
Uses DL and train generative statis
tical models on texts. AI helps
with automatic tagging and
allows users to search by
description, not keywords (13).
15. Several factors contributes to the success of AI, one has to find a way to exploit it
Factors contributingto AI development and ways to exploit it
(13) Jasnam S. Sidhu, and David Moloney, Price Waterhouse Coopers Presentation at Digital Catapult’s event, London, September 2016
Factors moving AI forward How to benefit from AI (13)
Ø Progress in technology;
Ø Growing investments;
Ø Interest and resources of the leading corporations;
Ø Cross-industrial and cross-functional character.
Ø Track new companies / products in AI;
Ø Define clear priorities on what to look at;
Ø Develop a relevant strategy;
Ø Build a relevant talent pool;
Ø Experiment with AI.
AI is here to stay, paying attention to it is curtail for success of a business
16. Thankyou!
Please feel free to let me knowif youwouldlove to knowmoreaboutAI
applications,marketsandinvestmenttrends
and/or
youare developinganAIstartup
pz@flintcap.com
checkfor insightsandupdates