How to Troubleshoot Apps for the Modern Connected Worker
AI in the post-COVID era
1. Alex Honchar | Neurons Lab LTD | 2020
AI in the post-COVID era
what broke, what stays, what appeared
2. 🚀
Entrepreneur
creating and selling AI
solutions worldwide
"
Practitioner
7 years building data-
driven products
🎯
Educator
university lecturer, events
speaker and blogger
About me
I am teaching machines
3. About Neurons Lab
Uninterrupted
energy production
20% less maintenance fees
Explainable AI
in Finance
10% higher Sharpe ratio
Human-free
merchandising
5% higher sales and margins
Fully distributed AI consulting boutique
4. Plan for today
18:00-19:30
• General AI landscape
what worked well?
• AI applications emerged in crisis
what broke? what has to be fixed?
• Future AI applications
what we are looking into?
6. Value chain
what this AI actually does?
Project Create Promote Provide Prevent
Retail
Anticipate, predict
and find sources ->
launch new ideas
Automate, augment
and optimize human
roles -> better
processes
Advertisement, sales
and pricing -> drive
the revenues
Enhanced user
experience,
minimized human-
based support
Risk management,
radical negative
events prevention
Wealth
Management
Utilities
Healthcare
7. Value chain
what results it delivers?
Project Create Promote Provide Prevent
Retail
1–2% EBIT
improvement
30% reduction of
stocking time
30% online sales
increase
5% higher margins
with merchandising
50% less burglaries
and stealing
Wealth
Management
More accurate
predictions
9%+ outperform
classical funds
General sales
increase
$2.2T
roboadvisers
market
20% less
unprofitable deals
Utilities
Cuts up to 10%
electricity usage
20% energy
production growth General faster and
better support and
delivery
$10–$30 savings on
monthly bills
Healthcare
$300 billion possible
savings in the United
States
Up to 2% GDP
savings for oper.
efficiencies
5–9% health
expenditure
reduction
0.2–1.3 additional
years of average life
expectancy
8. Value chain
what technologies are used?
Project Create Promote Provide Prevent
Retail
Time Series
Forecasting
Optimization,
Robotics
Recommender
Systems and KYC
Technologies
Computer Vision
and
Natural Language
Processing
Computer Vision
Wealth
Management Optimization and
Reinforcement
Learning and
Robotics
Anomaly detection
Utilities
Healthcare Computer vision, NLP, DSP
Generative
modeling
10. Affected industries
highest probability / least probability to default
Airlines 23.16%
Casinos & Gaming 16.46%
Leisure Facilities 19.74%
Auto Parts & Equipment 18.81%
Oil & Gas Drilling 22.5%
Property & Casualty Insurance 1.06%
Health Care REITs 0.52%
Multi-line Insurance 0.92%
Life & Health Insurance 1.00%
Industrial REITs 0.55%
12. Value chain affections
where AI broke here?
Project Create Promote Provide Prevent
Retail
Demand and Supply
patterns changed
Robotics here will be
very handy
People simply buy
less now
Shopping
experience changed
Facial recognition
doesn’t work
anymore
Wealth
Management
Forecasting doesn’t
work at all
The stock market is still alive, people are still trading
How to survive next
market crash?
Utilities
Predictive maintenance to avoid too much
of the human work
Do you really need to promote utilities
differently? The support is online already
What if all factory
workers get sick?
Healthcare
Doctors are not looking to be replaced by
AI here right now
Don’t see how AI will
help to sell here
How to scale
services?
Preventive medicine
with all the data we
have
14. Nowcasting
• Retail:
supply-demand price elasticity models
“broke” due to completely new market
situation, new online-learning models
are going to replace them
• Wealth management:
econometric models for price
prediction “broke” as well, alternative
data analysis will be a key for
nowcasting.
instead of forecasting
15. Remote operations
• Augment, not automate
- predictive maintenance: a way to reduce amount of time spent by workers
checking physical objects
- telemedicine is real and yes, bots can do the primary asking
- robotics has to arrive sooner, even it’s not ready yet
• Create new work places:
- each AI solution frees routine work of one type, but creates more of other types.
Industrial revolutions give more jobs than take away
work from home or at least less physical presence at the object
16. Public safety
• Public surveillance:
- are you far enough from others?
- are you wearing a mask?
- what temperature do you have?
• Workspace safety:
- are you coughing?
- are you wearing protective suits?
- are you washing your hands?
aka hello to the totalitarianism in the Western World
19. Conclusions
• Investors:
Social distance hole will be filled with the machines. Let them to help people to get rid of
the manual work and use their data to live longer.
• Entrepreneurs:
Build “nowcast” analytics on every step of the value chain: from business predictions to
executions
• AI experts:
Learn how to create the value for the guys above, it’s your time now!
20. ping me for details and to collaborate
alex.honchar@neurons-lab.com
21. Resources used
in this presentation
• McKinsey AI Report https://www.mckinsey.com/~/media/mckinsey/industries/
advanced%20electronics/our%20insights/
how%20artificial%20intelligence%20can%20deliver%20real%20value%20to
%20companies/mgi-artificial-intelligence-discussion-paper.ashx
• Affected industries https://www.spglobal.com/marketintelligence/en/news-
insights/blog/industries-most-and-least-impacted-by-covid-19-from-a-
probability-of-default-perspective-march-2020-update
• Demand pattern changes: https://blog.griddynamics.com/rapid-response-to-
covid-19-supply-chain-and-market-shocks/
22. Resources used
in this presentation
• Nowcasting vs Forecasting: https://papers.ssrn.com/sol3/papers.cfm?
abstract_id=3562025
• McKinsey on COVID: https://www.mckinsey.com/business-functions/risk/our-
insights/covid-19-implications-for-business
• AI computing prices plot: https://theassets.ru/
• Proprietary Neurons Lab documentation and knowledge base