Unstrukturierte Daten strukturieren – der Anfang von Analytics und Digitalisierung; Wie mit Legacy-Umgebungen in der „neuen Welt“ umgehen?; Mit Automatisierung und APIs zur agilen IT-Infrastruktur
5. Page 5
2003
COMPANY TIMELINE
SWISS POST SOLUTION
Swiss Post Solutions | 06.06.2019 | V1 | public | DATA 2019 Storage & Analytics Fachkongress
End-to-end productive
process automation
1st RPA-Robot live
BpaaS
RPA
Test Automation
AI-powered
Email Client
2013
2017
2019
2003
Artificial
Intelligence
End-to-end test process
automation
Document Input Processing
Smart Process Discovery
1st AI live case
Sentiment
Analysis
Data
Analytics
Speech-to-speech
6. 60% of customer dissatisfaction
originates within back office
inefficiencies
Backing up the Digital Front:
Digitizing the Banking Back Office
Capgemini study 2013
«A GOOD BACKOFFICE
MAKES CUSTOMERS HAPPY»
10. Page 10Swiss Post Solutions | 06.06.2019 | V1 | public | DATA 2019 Storage & Analytics Fachkongress
UNSUPERVISED CLUSTERING & PREPROCESSING
HOW TO GET THE RELEVANT DATA
REGEX REJECTION 9’669
After 3-step pre-processing ~6’000 relevant mails left
TOTAL 66’466 MAILS
37’094 German
9’669
LANGUAGE DETECTION
REGEX-BASED
REJECTION
CLUSTERING~6’000
12. Page 12
PROCESS RE-ENGINEERING
AUTOMATE PROCESSES
INCREASE OUTBOUND CAPACITY
Swiss Post Solutions | 06.06.2019 | V1 | public | DATA 2019 Storage & Analytics Fachkongress
AI PROJECT @CONTACT CENTER
INITIAL SITUATION
INCREASING MAIL VOLUMES – COSTLY
KEYWORD-BASED CLASSIFICATION
INCONSISTENCY AMONG OPERATORS
CHALLENGES GOALS
13. Page 13Swiss Post Solutions | 06.06.2019 | V1 | public | DATA 2019 Storage & Analytics Fachkongress
AI PROJECT @CONTACT CENTER
INITIAL PROCESS SITUATION
Unwanted Mails
Lots of unwanted emails (e.g.
LinkedIn requests) end up in inbox.
On the other hand, ~ 900 customer
requests/ year categorized as spam
and deleted without interaction.
Private / Commercial
Although required, current system
cannot distinct between private and
business customers
Stamps Order
Simple order forwarding process,
predestined to be automated
14. Page 14
AI PROJECT @CONTACT CENTER
AUTOMATION RESULTS
Swiss Post Solutions | 06.06.2019 | V1 | public | DATA 2019 Storage & Analytics Fachkongress
15. Page 15Swiss Post Solutions | 06.06.2019 | V1 | public | DATA 2019 Storage & Analytics Fachkongress
BPO SERVICE
Our solution is designed to automate simple queries
and to assist employees with complex business
processes through preparing unstructured data.
16. Page 16Swiss Post Solutions | 06.06.2019 | V1 | public | DATA 2019 Storage & Analytics Fachkongress
BPO SERVICE
PROCESS MODEL
1) FIRST LIVE PROCESS
BUILD FRAMEWORK
FAST ONBOARDING OF 1ST PROCESS
2) PROCESS AUTOMATION
ENHANCE E-MAIL ROUTING
PROCESS SIMPLIFICATION
3) EMAIL CLIENT
REDUCTION OF AHT
CUSTOMER SATISFACTION
PHASE1
PHASE2
PHASE3
17. A BALANCING ACT BETWEEN
DIGITAL EXPERIENCE AND
LEGACY SYSTEMS
19. Page 19
ROBOTIC PROCESS AUTOMATION
SHARED SERVICE CENTER
STRUCTURED DATA
EmployeeMobile App
Paper Form
SPS Document
Processing
Automated
Data Verification
Automated
Back Office
Data entry
INITIAL SITUATION
Employee
Paper Form Manual
Back Office
Data entry
SPS SOLUTION
STRUCTURED DATA
2
DAYS
30
DAYS
Swiss Post Solutions | 06.06.2019 | V1 | public | DATA 2019 Storage & Analytics Fachkongress
20. Page 20
ROBOTIC PROCESS AUTOMATION
SHARED SERVICE CENTER LIVE PROJECT: RESULTS
Swiss Post Solutions | 06.06.2019 | V1 | public | DATA 2019 Storage & Analytics Fachkongress
< 1 year
ROI
70%
Reduction
Handling time
40’000
Transactions
per month
82%
87% 89% 90% 93%
Start
(CW27/2017)
First iteration
(CW35/2017)
Second
iteration
(CW42/2017)
Third iteration
(CW08/2018)
Fourth iteration
(CW24/2018)
Success rate
22. Page 22
E-mail
Swiss Post Solutions | 06.06.2019 | V1 | public | DATA 2019 Storage & Analytics Fachkongress
YOUR CONSULTANT
Lukas Hebeisen
Head of Solution Development & Digital Transformation
Post CH Ltd
Swiss Post Solutions
Pfingstweidstrasse 60b
8080 Zürich
lukas.hebeisen@swisspost.com