2. About Me
CEO @ API Garage
Technology Mentor @ Biomedical Zone ( Ryerson University &
St. Michael’s Hospital )
Founder of Applied AI Toronto Community (
https://www.meetup.com/Applied-AI-Toronto/ )
5. Big Data-berg
AI is required to
explore hidden
potential within
that data.
Only 0.5% of data is being used
for analytics. (Source: IDC)
Evolutionary Strategies
Deep Learning
Churn Prevention
Anomaly Detection
Recommendation Engine
Sentiment Analysis
Predictive Analysis
7. Built an Automated Claim Resolution Bot -
Jim
18 anti-fraud algorithms
3 seconds claim resolution
0 paperwork
Automated money transfer, Great user
experience.
11. Business Problem Identification
Qualitative Leads
● Look for processes that are:
○ Repetitive
○ Well documented
○ Data driven
● 1s Rule : “Given all the data, can an expert take the ‘correct’ decision within a second”
Improved Targeting Personalized Experience
12. Data Gathering & Preparation
Data can be Structured or Unstructured. Data annotation services are available for unstructured
data. Example: Mechanical Turk by Amazon.
Partners Data
● Affiliates
● Service providers
● Data partnerships
Internal Data
● ERP/CRM Data
● Marketing
● Sales
Public Data
● Paid / Free
● Scrapped
● Next slide
14. Model Training & Deployment
Most of the common models are available Open Source.
● For this problem we can use simple prediction service using H2O or Electric Brain or
TensorFlow.
● Depending on your current environment, you will select your tools.
15.
16. Key questions to ask your team:
● What is the frequency and volume of data updates?
● How accurate the model will get if you add new data?
● What infrastructures is required to support frequent updates?
● What is the current infrastructure?
● How it will be integrated back in the system / product?
Maintenance & Enhancements
In Our Case:
● Update model every 3 months.
● Look for continuous open data to include in the model
● Considering scrapping the digital dust on the internet
17. Wealth Management - User Segmentation
Artificial
Intelligence
Use clustering algorithms to group
users and find out micro-segments
within the current user base.
Data
Platform
New User
Segments
Digital
Data
Financial Institution
Data
Brokerage and
Custodian Data
New Revenue and
Product Opportunities
Increased Customer
Loyalty
Customized Reports
and Engagement
18. 18
Build AI use cases for
your organization
Use AI Design Sprint or many other
methodologies to build AI use cases.
1
Clean up and annotate
your data
Collect, clean, and enrich your data. If you
are working with a third party vendor make
sure legality is covered.
2
Start experimenting
Experiment internally, first. As leaders in
your firm, it’s your duty to change this.
3
Next
Three
Steps
19. Questions?? How can I help?
Chinmay Patel
CEO @ API Garage
(647) 878 4217
cpatel@apigarage.com
www.apigarage.com