Upload
others
View
3
Download
0
Embed Size (px)
Citation preview
s
ssCopyright © 2018 NTT DATA INTRAMART CORPORATION
Business innovation of BPO
realized by Task Center and AI and Rule Engine.
President & CEO
Yoshihito Nakayama
NTT DATA INTRAMART Corporation
Copyright © 2018 NTT DATA INTRAMART CORPORATION 1
A demand for BPO has greatly increased, but there is room for various improvements in the productivity.
The improvement of business productivity using AI and rule engine.
1. “Task Center” has a mechanism of collecting various tasks to be tackled automatically from existing systems and aggregating them on the screen.
2. “Dash-board” has a mechanism for allocating tasks to team members optimally according to work volume and skills through business visibility.
3. Method for visualizing the situation of complex on-site operations. Possible to construct “digital twin” of on-site business processes in any business.
Abstract
Copyright © 2018 NTT DATA INTRAMART CORPORATION 3
Collect information from various systems and applications on behalf of users.
Create task lists automatically.
Users can understand what they should do by using “Task Center”.
Concept of Task Center (1)
Copyright © 2018 NTT DATA INTRAMART CORPORATION 4
Collect tasks of the entire team as well as individuals and automatically distribute tasks to members optimally.
Concept of Task Center (2)
Task Center
SFDC
SAP
O365
Groupware
Collection of tasks / Optimal allocation
Copyright © 2018 NTT DATA INTRAMART CORPORATION 5
Screen of IM-Task Center
Estimation of completion date and time accurately
Task Center
Task occurrence date
Start date Due date Estimated time
PersonPriority StatusWork Summary Forecast completion
date
Recommendation
● performance ○ Operation leveling ○ Cost
Collection of tasks automatically
Add line Delete line Recalculation Reflection of recommendation
Recommendation of allocation by AI or Rule
Engine
Copyright © 2018 NTT DATA INTRAMART CORPORATION 6
Issues that can be solved by Task Center
Take a lot of time to manage members and allocate tasks
Allocation of tasks
manager
members
Not know how to allocate unexpected work.
Not know the operational status of members.-> Can not grasp whether to meet the deadline
Not know the priority among member’s own tasks
Copyright © 2018 NTT DATA INTRAMART CORPORATION 7
Benefits of Task Center
Automatic Allocation
Task Center
- Integrated management of various tasks- Less missed tasks and easier management of progress
1) Visualization of all tasks Support for optimal task allocation based on member productivity and workload forecastusing AI & Rule Engine
2) Efficiency of allocation work
Forecast of future work such as overtime and cost using information such as work volume and member skills as input data
3) Visualization of future forecasts
Copyright © 2018 NTT DATA INTRAMART CORPORATION 10
Mechanism of AI prediction and optimization
Task occurrence
Worker information
Predicational task
Actual Task
Automatic allocation
Task type estimation
Task standard work hours estimation Standard work hours
Type
Task information
Allocation of task
Task registration
Personal productivity estimation Productivity
Task DB
Prediction of task occurrence
NumberType
Standard effort
Optimal task allocation based on member productivity and workload forecast
・Prediction of task occurrence in future
・Prediction of standard work hours
・Task type classification
・Estimation of individual productivity
Copyright © 2018 NTT DATA INTRAMART CORPORATION
assign
11
Optimization of task assignment
Solutions for conflicting goals & Restrictions.
Minimization of work time / save of labor
costs / meet of deadlines / Operation
leveling
Compliance with work regulations (labor-
management agreement / legal regulations)
… …
TaskMember
Due dateSchedule
attempt
Goal(Solved solution)
Final decision can be made with human judgment using Task Center
Copyright © 2018 NTT DATA INTRAMART CORPORATION 12
Predicting future task volume and member load situation from past task information.
Prediction of task volume
Prediction with ensemble of regression model and time series model
Copyright © 2018 NTT DATA INTRAMART CORPORATION 13
Prediction of task volume
Estimation of
work requests
number for each
task type on a
daily basis up to
one month.
Copyright © 2018 NTT DATA INTRAMART CORPORATION 14
Business system Task Center
Agent
Connector
Connector
Connector
Crawler
Acquisition of dataConvert to standard format
Predict
Summarize actuals for display Dash-Board
Task Center
Mechanism to collect task information from business system
Machine Learning
Person in charge
ManagerAcquisition of data
<Information to be acquired>・Operating system name (application)・ What task・ Who・ When (start to end)・ Project・ Work volume (frequency)・ Data volume (cumulative volume)・ Master to access
Copyright © 2018 NTT DATA INTRAMART CORPORATION 17
Dash-board of business process(1/2)
Future business volume forecast
Drill down analysis
Copyright © 2018 NTT DATA INTRAMART CORPORATION 18
Dash-board of business process(2/2)
[Simulation Policy]● Cost saving ○ performance ○ Operation leveling ○ Current status
Simulation per policy
[Display Condition of Dash-board]2018/9/1 〜 2019/2/1 ● Incompletion Task
Analysis of past business conditions
Dash-board portal Dash-board portal
Number of tasks and volumes per project
Dash-board portal
Copyright © 2018 NTT DATA INTRAMART CORPORATION 19
Demonstration of “Dash-board of business process”
Copyright © 2018 NTT DATA INTRAMART CORPORATION 20
Analysis of all of the work log including RPA
Number of orders received
Number of complains
Copyright © 2018 NTT DATA INTRAMART CORPORATION
• How can you easily grasp the real status of field work?
Copyright © 2018 NTT DATA INTRAMART CORPORATION 23
Technology to identify business bottlenecks from existing system logs and extract processes automatically
Improvement of the current work quickly is possible while leaving the existing system
Process mining
Data Logs
Process mining
Visualization of process logs
Copyright © 2018 NTT DATA INTRAMART CORPORATION
24
Logs can be collected even in environments that do not use the system (1/3)
1. Confirm the digital memo
2. Touch the digital memo and employee ID card
4. Confirmation of update of digital memo
In ActionProcess No.2Kelly
3. Tap the start of work
Process No.1Process No.2Process No.3Process No.4Process No.5 Start
Start
Start Done
Done
Done
Done
Cancel Update
Already
AlreadyAlready
ID 1234567 Kelly
Customer: NTT DATA INTRAMARTReception Date: 14:04 01/12Case Name : Order #1Number od case : 120Due date : 14:53 01/12
Copyright © 2018 NTT DATA INTRAMART CORPORATION 25
Logs can be collected even in environments that do not use the system (2/3)
Simulation of optimization on
digital twin
Process Design
Visualization
Monitoring
Copyright © 2018 NTT DATA INTRAMART CORPORATION
Touch!
Get the situation of work in the field in
various ways
How to create Digital twin of business process
Transformation
BPM
Automatic recording of who, when, what and how long
Logs DB
Process Mining
Dash-Board
Automatic generation
Visualization and processing of work
Copyright © 2018 NTT DATA INTRAMART CORPORATION 28
Case studies
ID Sector of customer Contents
1 BPO Utilization of Task center in BPO work
2 Call center Utilization of task center in Call center work
3 Trading company Utilization of task center in office work
4 Construction company
Business bottleneck search at construction site
5 Manufacturing company
Use Dash-boards in manufacturing productivity analysis
6 Financial company Business bottleneck search at bank window
Copyright © 2018 NTT DATA INTRAMART CORPORATION
Case studies currently underway
• Automating non-routine work
Business operation while AI and humans complement
each other
Routine work
Business Operation according to the rules
by the rule engine
Non-routine work with experience in the past
Make decisions based on AI recommendations
Non-routine work with no experience in the past
Flexible operation by humans
Copyright © 2018 NTT DATA INTRAMART CORPORATION
The best response to non-routine work can be shared with all members
Point 1
CaseManagement
The best response to non-routine work
Best choice for similar non-routine work
Change the handling of frequent non-routine work to automatic routine work
Point 2
CaseManagement
Similar non-routine works occur frequently
Defined as a new automated routine work
Copyright © 2018 NTT DATA INTRAMART CORPORATION
“Case Management” orchestrates not only BPM but also RPA and human operation based on the judgement of Rule engine and AI.
Case ManagementRule
Engine
AI
BPM RPA Human etc.
Transform non-routine work into routine workfor full automation
Microflow
RPA
Human workflow
BPM
Copyright © 2018 NTT DATA INTRAMART CORPORATION 33
Conclusion
Centrally management of tasks using the Task center.
Creation of digital twin using process mining and other tools even at sites not using the system.
Consolidate team tasks and allocate them optimally.
Visualization of business conditions and forecast on a dashboard.
Introducing case studies and current challenging.