Upload
others
View
20
Download
1
Embed Size (px)
Citation preview
Artificial Intelligence: Enhancing Due Diligence &
Prospect Research
David Payne – Client Success & Engagement Manager, Exiger
Wouter Servaas – Prospect Researcher, University of Sheffield
3 Confidential All rights reserved. Copyright © 2017. Not for redistribution
Agenda
Introduction to AI – understanding cognitive computing and natural language processing
How AI can aid prospect researchers in due diligence?
Live demonstration
AI in prospect research – an adopter’s experience
Q&A
4 Confidential All rights reserved. Copyright © 2017. Not for redistribution
Data Protection and Compliance
General Counsel Statement:
Exiger infrastructure and data is co-located in redundant SOC 2 Type II certified, Tier 3 data centers. The data center operators only provide power, cooling, physical security, and carrier network connections, Exiger owns and operates all hardware and software
We are reviewing all data transfer compliance and we believe we are in compliance with GDPR standards now but we definitely will be before implementation in May 2018.
Thus far Canada has been white listed by the EU data authorities as meeting adequate PI protection standards: http://ec.europa.eu/justice/data-protection/international-transfers/adequacy/index_en.htm
DDIQ –
Automated Due
Diligence
6 Confidential All rights reserved. Copyright © 2017. Not for redistribution
6
DDIQ Use Cases
Major donor research
Donor prospecting
Corporate relationships
Honorary degree nominations
7 Confidential All rights reserved. Copyright © 2017. Not for redistribution
New Paradigm in Due Diligence
2000s 2010s Today
Watchlist/
Development
Data
Aggregation
Automated
Due Diligence
8 Confidential All rights reserved. Copyright © 2017. Not for redistribution
Why Context Matters
Google: Consumer focus, too many false hits, often wrong context, shallow
Cognitive Adjudication Platform: Contextual analysis minimizes noise by navigating through the results as an investigator
Sample CAP-Generated Report Too many results
Wrong Subject
Wrong Subject
again
Right Subject
but not a risk
Wrong Subject
again
Search: Clean
9 Confidential All rights reserved. Copyright © 2017. Not for redistribution
How does DDIQ work?
Investigative &
Enhanced
Due Diligence
Real-World
Examples
Cognitive Computing
Platform
Report
10 Confidential All rights reserved. Copyright © 2017. Not for redistribution
Entity-Based Cognitive Computing Technology
Structured & Unstructured Data Acquisition Constantly growing list of sources
Aggregation of open web and premium sources
Proprietary Auto-Heal data connections
Comprehensive & Accurate Research Structured extracted information
95% of false positive removal
Easy access to source material
Profiles generated in < 5 minutes
DDIQ Cognitive Engine Natural Language Processing learns about the Entity
Assesses content and extracts information
Eliminates false-positives
De-duplicates content
Data Input Subject Information
W W W
Enriched seed data
11 Confidential All rights reserved. Copyright © 2017. Not for redistribution
DDIQ: Investigative Cognitive Computing
Natural language processing allows DDIQ to read, understand and piece together text. It discovers and follows leads found with the text similar to what a human researcher would do.
The Old Way The Way
12 Confidential All rights reserved. Copyright © 2017. Not for redistribution
DDIQ: Automated Due Diligence Reports
IPO
Screen
Rapid
Manual Diligence
Scope
Watchlist / Database
Screening Required for
global
compliance as
well as
operational &
reputational risk
reduction Red Flag Level I Level II
DDIQ
Risk
Profile
~ 10 Minutes 5 Business Days 7-10 Business Days 10-12 Business Days Bespoke
Profile Narrative Report
13 Confidential All rights reserved. Copyright © 2017. Not for redistribution
Cognitive Computing Research
DDIQ’s open web searches present information that may be removed regionally.
14 Confidential All rights reserved. Copyright © 2017. Not for redistribution
Found Hidden Search Results
Warning that the results have been removed due to EU law
Search results related to corruption have been removed
Not all results are available in English
Search results related to corruption are surfaced
Non-English results have been translated
Identified risks are categorized by event type
15 Confidential All rights reserved. Copyright © 2017. Not for redistribution
DDIQ Deployment Models
SaaS Cloud Per User Access
Batch Reporting Automatic Run of DDIQ Profiles
Account Refreshes/Remediations
Enterprise Integrated DDIQ Engine
Embedded in Client Workflow
Available Hosted, On-Premise,
And in Hybrid Deployments
Profile Monitoring Available for All Configurations
16 Confidential All rights reserved. Copyright © 2017. Not for redistribution
DDIQ
DDIQ Demo
17 Confidential All rights reserved. Copyright © 2017. Not for redistribution
DDIQ Client Success Story – University of Sheffield
Pain Points
Spelling of prospective donor names, especially when translated from languages with non-Latin alphabet
Use of aliases, married names, or maiden names
Prospective donors with generic names – how to weed out the false positives
Prospective donor organisations without websites
A world top-100 university, renowned for the excellence, impact and distinctiveness of its research-led learning and teaching, set out its own due diligence process to safeguard its reputational standing.
18 Confidential All rights reserved. Copyright © 2017. Not for redistribution
DDIQ Client Success Story – University of Sheffield
DDIQ Solutions
DDIQ takes into account various spelling options
DDIQ includes aliases, maiden names and married names in the search results
Being able to fill in DOB and alma mater narrowed down the number of search results
DDIQ provides overviews of board members / trustees, enabling a more in-depth due diligence check
A world top-100 university, renowned for the excellence, impact and distinctiveness of its research-led learning and teaching, set out its own due diligence process to safeguard its reputational standing.
19 Confidential All rights reserved. Copyright © 2017. Not for redistribution
Case Study: University of Sheffield
AI helps the University of Sheffield to streamline its due diligence process and increase the turnover rate of due diligence checks
The University was offered a donation by an organisation. The organisation’s name came up in online searches, but little was revealed about its internal structure. A due diligence check was needed to ensure there would be no reputational risk for the University in accepting this donation.
Artificial intelligence helps human researchers in creating the full picture. At the University of Sheffield, DDIQ has proven to be an excellent addition to existing due diligence practices, enabling the University to make fully-informed decisions on whether proposed donations can be accepted or not.
Challenge:
Solution:
20 Confidential All rights reserved. Copyright © 2017. Not for redistribution
Case Study: University of Sheffield
21 Confidential All rights reserved. Copyright © 2017. Not for redistribution
Q&A