Week 6 ESG - Jaffa.Net

ballooncadgeInternet και Εφαρμογές Web

31 Οκτ 2013 (πριν από 4 χρόνια και 13 μέρες)

118 εμφανίσεις

Alchemist


The Software for Norkom’s
Solutions


Enterprise and
Legacy

data

Algorithms for caching

and optimising

data harvesting

Apply

Intelligence

Information
Dissemination:

Deliver
Intelligence &

Information where
& when


it is needed

What is Alchemist
?

Multiple

Delivery

Channels

Notification

Technologie

Alchemist Design Principles


Modular in design and roll
-
out options



Robust & Open Technology Platform



Real
-
time and batch transaction monitoring



User
-
friendly, easy to use



Based on Web
-
enabled technology




Alchemist CRM


Intelligent Marketing Automation

Norkom’s View of CRM

Internal & External
Intelligence Gathering

Brand

Market

Customer

Channel

Product

i2i hub

Message Management & Notification

Business Rule Application

Opportunity Detection

Creative

Internal

External

Analysis

Predictive

Descriptive

RoI

Management

By

Intelligence

Proposition
definition &
authorisation

Above the line

Below the line

In
-
bound

Out
-
bound

Internal

External

Message Delivery

Media

Dealers /

Competition

Web /
Email

SMS

WAP

Billing

Direct Mail

Customer
Care

Service
Centres

Product & Service Design & Provision

DOL

WAP

FoneCover

SMS

SpeakEasy

Personal

Business

Handset

Where is Value Added and When?

Internal & External
Intelligence Gathering

Brand

Market

Customer

Channel

Product

i2i hub

Message Management & Notification

Business Rule Application

Opportunity Detection

Creative

Internal

External

Analysis

Predictive

Descriptive

RoI

Management

By

Intelligence

Proposition
definition &
authorisation

Above the line

Below the line

In
-
bound

Out
-
bound

Internal

External

Message Delivery

Media

Dealers /

Competition

Web /
Email

SMS

WAP

Billing

Direct Mail

Customer
Care

Digifone
Centres

Product & Service Design & Provision

DOL

WAP

FoneCover

SMS

SpeakEasy

Personal

Business

Handset

Identify new Value
propositions based on
enriched
customer/prospects
knowledge and market
profiles

Provide a consistent message
accross channels


Propose the right product

@ rigth time

@ the right customer/prospect

Learn from
experience and drive
Business process

Full Leverage requires
ability to drive
product/services design
according to findings

Analysis and Customer Lifecycle

Customer Relationship Management
-

Micro

Account
Application

Monthly
Statement

Rate
Enquiry

Upgrade

Complaint

New Job

Marriage

Children

New Home

Retirement

Competitor
Promotion

Competitor
Promotion

Usage

Exit

Customer
entry

Customer
Exit

Customer
Management

Credit

Scoring

Cross
-

Sell

Fraud

Profitability

Churn

Winback

Promotion

Functional Architecture





Alchemist Intelligence

Alchemist Intelligence Workbench
Features

Data Mining Suite

DESCRIPTIVE



Alchemist offers a
comprehensive range of
descriptive techniques including:


Association Rules


Kohonen Neural Network


K
-
means


Decision Trees


eg C4.5
(This can be viewed as
both a descriptive and
predictive technique)


Data audit tools


PREDICTIVE



Alchemist’s strong portfolio of
predictive techniques includes:


Robust Regression


Naïve Bayes


Back Propagation Neural
Network


Decision Trees eg. C4.5



Predictive Modelling


Outputs of a Predictive Exercise


Scored list






Model Contributions






Individual Contributions

T O’Reilly Ltd
: Score = 0.9 from
turnover
=high and
ave_contract_class
=
C

AR Building contractors Ltd
:
Score =
0.9 from
turnover
=
low

and
ave_contract_class
=
R

Customer

Information

Tenure

Billing

Location

No. of

Intl

Occup

Dealer

Usage

Information

No. Premium

Rate

No. of

calls

No. of

dropped

No. of SMS

Complaints

No.

Contacts

Interaction

Information

First/Last

Contact

Campaign

Letters

Call

Centre

Inbound

Dealer

Outbound

Intelligence

Workbench



Dependants

Use of information you have (today and in the future)

Price Plan

Roaming

Product

Information

Pre/Post

Dealer

Handset

Insurance

Corporate

Discount

Upgrades

supplemental



Alchemist Marketing Automation



Marketing Automation



The Automation of the marketing task of managing intelligent contact with
Customers



Closed Loop from Customer data through Modelling and Insight to Customer
Contact



Customer Intelligence Capability: Modelling, Profiling & Model Management



Campaign Management: Batch, Event Driven and Real
-
Time contact with
customers through all channels (Phone, DM, E
-
Mail, Call Centre etc.)



Benefits of Marketing Automation

Drive up response rates to campaigns


Timely and relevant customer offerings


Intelligence driven customer contact strategy


Easy selection of the right customer for right campaign


Better models driven by central contact and response repository


Optimise message and channel selection


Run 100 focused campaigns not 10 broad
-
brush campaigns

Drive down the cost of acquisition and conversion


Capture and report on campaign effectiveness


Optimal re
-
use of effective campaigns (learn by doing)


Reduce waste by contacting only ‘good’ prospects








Alchemist Marketing Automation Features


Campaign Definition


Automatic customer list generation based on business theme
being addressed
(
e.g.
Cross
-
sell)

taking in
to

account


volumes, profiles, campaign content

and

channel



Automatic Campaign Execution


Execute across multi
-
channel environment while:


Achieving channel optimi
s
ation


Include necessary authori
s
ation logic



Automatic Campaign Reporting


Review predicted vs. actual campaign performance


Understand customer to channel preference


Build product to customer profile by channel learning from


previous campaigns



Operational
CRM
Platform




Enterprise
Automated
Marketing
Platform






Critical
Business
Issues








Call Centre



Direct Sales

Channel


Broker /

Correspondent
Channel



Direct
Marketing



Web


mCommerce

-

PDA, SMS,

WAP, Pagers






Customer
and Third

Party
Information


Prepayment
Risk


Customer
Retention


Cross
-
Sell


Life
-
time
Value


Channel
Optimisation


Customer
Acquisition


Credit
Risk


-

Solimar/Base100
-

DataQuick

-

CSW
-

Equifax

-

MRAC
-

TRW

-

MIC
-

Experian



-

Mortgage Servicing Database

-

Retail Bank Database

-

Origination Database

-

Marketing Database


-

Acxiom InfoBase

-

MITI


Predictive Applications

Intelligent Information Workflow

Campaign Management

Other Data Sources


-

House Price Indices


-

Pricing & Credit Data


Customer
Database



Alchemist
Risk & Fraud Management



Integrated Solution for Financial Crime











Alchemist for AML modules for


Case Management


Scenario detection engine


Customer Profiling engine


Advanced Name matching with prepopulated
data dictionary and language rules


List management


Reporting and audit


Security and user administration


Real time and batch operation


Standard Interfaces



Automated 3
rd

party list integration


E.g. World Check


Provision of World Check list of PEP’s and
Sanction lists etc.

Alchemist Configuration

!

Source

Transactions

& Data


Alchemist

Data Model



Business Rules


Unusualness Tests


Predictive Modelling


Network Analysis

Compliance

Workflow


Analysis &

Reporting



FIU Email & Fax

Dashboard

Manual Reporting

Watch Lists


Watch List Monitoring


Watch List Rules

OFAC,

BOE

EU

World
-
Check etc

The Alchemist HUB

Alchemist Fraud Functional Architecture

Name

Matching

Rules

Engine

Case

Management

Analysis

&

Reporting

Profiling &
Segmentation
Engine

Intelligent Agents

Alerts & Notification

Portal


Knowledge
Base

Alchemist AML

Alchemist Name Matching


Data dictionary per country


Multiple character sets


Consolidated list management


From multiple sources


Sanction lists + PEP’s


External and internal lists


Automated list update and refresh


Fuzzy logic search capability


Support for multiple jurisdictions


Flexible audit and reporting capability


Integration with Alchemist Case Development, Workflow and reporting


Search & Matching Ability


peoples names


account or compound names


delivery or postal addresses


additional attributes


Covered by Population Rules


Covered by Language Rules


Covered by Character set Rules


regardless of

spelling variation, phonetics, abbreviation, nicknames, missing
or extra words, word sequence variation or data quality ….


with
controllable

performance and response times

Rules Engine


Detailed Rules


insurance,
banking, investments


Flexible


Detect unusual behaviour for
accounts and accounts in
segments


Robust, scalable


Rules maintained by MLRO, not
IT
, easy to use


Operate automatically

Rule
Triggers

Early policy cancellation in unusual
circumstances

Cash Deposit, many ATM
withdrawals

Account

Transaction


Profiles

Suspicious
Cases

Case Management

Automatic Prioritisation


Rules that are triggered or
predictions above thresholds
create suspicious customer
records.


Detailed customer profile and
transaction history sourced


Several states:


Triggered, Monitored, Reported,
Clear, Investigation


All cases are prioritised by
further rules:


Historical activity/cases


Severity of rule Predictions


Full audit trail and security

Segmentation and
Risk Scoring


Based on continuing behaviour, behavoural profile models
are built/rescored for each account, trading house and/or
company


Notification of ‘Unusual’ account / trading behavour


Client behavioural clustering / movement


Increase in business knowledge (client, house, market)


Early indication of equity trading outside ‘normal’ price
bands



Based on historical records, predictive models for each
segment are built


Propensity to exhibit
market abuse
activity scores
created for each account


Allow for stastical tuning of business rules


Fraud Modules



Card Fraud


Debit Cards


Credit Cards


Skimming


Kiting


Devices


Branch Fraud


e
-
Banking Fraud


Identity Theft


Access Behaviour


Device Analytics


Tax Fraud


VAT Fraud


Insider trading


Market Timing


Loan Application Fraud


Loss recovery


Internal Fraud





Business Intelligence and

Performance Management



What are we trying to achieve?

Balanced

Score Card

Customer

Report Card

Cross

Functional Analysis

INTELLIGENCE

DATA

On
-
Line Analytical Processing

On
-
Line Transactional Processing

Subject Area Databases

Customer Segmentation

Micro
-
Marketing

Sales

INFORMATION

Marketing

External Customer Data

Engineering……

Financial Data

We want to put in place a
n
environment that facilitates
proactive, strategic,

business decisions

at the
appropriate level within
Banks abd Telco’s
by
exploiting the

data contained in

their
Information

Systems

Event
-
based Reporting / Decisioning

Alchemist

Agents

Analysis

& Reporting

Portal


&

KPIs

Alchemist

Rules Engine

Alchemist

Intelligence



Simulated Business Modelling



“What
-
If” Analysis



Forecasting




Executive Alerting



Survey distribution



Survey Collection



Alert Escalation

Balanced

Scorecard

Datamart



Personalised



Digital Dashboard



Dedicated KPI Module



End
-
user Query



Production Reporting



OLAP Analysis



Data Collection



Transformation



Data Loading

Visualisation

An example showing some
chart type variations

Example of an Architecture

System Architecture

Technology Options

Technology

Technology Options

Web Server

IIS, Sun ONE, Oracle 9I AS Apache,
Websphere, Weblogic

AML Application server

Apache Tomcat, Websphere,
Weblogic, Sun ONE, Oracle 9I AS

AML Database

Oracle, DB2, SQLServer, Teradata,
any JDBC compliant relational
database

Operating system

SUN 2.6, 2.7 .28, IBM AIX 4.3 & 5.1,
HP 11 & 11i, Microsoft 2000, NT 4

Client platforms

Standard Win98, WinNT, Windows
2000, Win XP desktop, IE5.5

Alchemist standard options


Supported Hardware




Supported Operating Systems




Supported Databases


Hub based approach

External Data

Centralised AML Server


Web Server

Operational Systems

Centralised Data Warehouse