Intelligent Process Management & Visualization Technologies

Intelligent Process Management & Process Visualization
TAProViz 2014 workshop
Presenter: Dafna Levy

The Topics
Process Visualization in Priority ERP
►Planning
►Execution
►BI analysis (Built-in)
Discovering implementation & operational issues (with Disco)
Click, View & Do! - Process Intelligence (PI) in action
Disco on the production floor - a hot opportunity
Q & A

A Hedgehog in his mating Season*
A Process Model
Encrypted mission of a Mosad agent
None of the above
Possible Answers

The right Answer: A Process Model (in Priority ERP)

The BPM Flowchart in Priority ERP
Built-in mechanism to define and control process steps
The model includes statuses and paths between them - the desired workflow

Business Rules in Priority ERP
A tool to define business goals for a business process
Enables to define time-based rules & alerts
Analysis of all deviations from rules is not enabled in the ERP system

Handling a Sales Order
Display of the sales orders screen in Priority ERP
When updating a status or an employee in a sales order, an event record is created in a sub-level form
The history of statuses data (i.e., event log), together with business data can be easily retrieved and exported from Priority for process analysis

The Sales Manager Dashboard in Priority ERP
The dashboard displays information about open sales orders
Currently, there is no way to discover inefficiency, bottlenecks with reports generated by the ERP system

Improving the Implementations of Priority ERP with Process Mining Technology

Some Questions that can be Answered
Can we visualize how our BPM flow chart is actually executed?
Are there any bottlenecks in the process?
Are there problematic paths taken between statuses?
What about fulfilling SLA conditions?
Which and how many employees deviate from rules?
Why are there performance differences among branches?

Process Mining Started in Eindhoven

Fluxicon: The Parents of Disco

Fluxicon’s Academic Initiative
Academic institutes in Israel are surrounded with rectangles

The Project: Analyzing Purchasing Process in Priority ERP

The Purchasing Process
Three separate sub-processes
Three different documents in Priority ERP (i.e., three different log files)

The Project Goals
Discovering how the Purchasing process is implemented & executed in Priority ERP
►End-to-End (a must!)
►Sub Processes
Identifying implementation & operational issues
Present results to top management
Define process targets and metrics

Some of the Analysis Requirements
Various visual maps of the common process
 Similarities and differences from the ideal process
Durations of cases & statuses
Bottlenecks & delays
Work loads on the various process participants (buyers)
Rework
Undesired paths

The Process Model in Priority ERP

The Planned Workflow
Bizqagi process modeler was used to design the process flow according to the data & information received from the customer

The End-to-End Process in Reality
Three event logs were joined with QlikView
The process model was discovered with Disco

A Live Demo of Process Discovery With Disco
Handling of the Vendor Quotes in Priority ERP*
*Anonymized data are used for the demo

The Vendor Quotes Event Log in Excel
The mandatory data labels are marked in red
The extended data labels are marked in green

Importing the Event Log to Disco

The Process Map View
The Vendor Quote process model which was automatically generated by Disco

The Statistics View
General & Statistical info about the log is displayed in charts and tables

The Cases View
Detailed view of the cases and events
Grouping by variants

Extending the Analysis
The type of the Vendor Quote approval depends on the purchase quote price
A more focused analysis was required according to Vendor Quote prices
The event log was extended with the relevant data in order to focus the analysis by these attributes
Type of Approval
Price Range ($)
Automatic Approval
<30,000
One Signature
30,000 – 100,000
Purchasing Committee
>100,000

Using Filters
The Attribute filter in Disco was used to analyze the cases according to price ranges
Here, the highest price range is selected

Discovering Deviations
After applying the filter, It was obvious that there are vendor quotes that were not approved as expected (i.e., by the purchasing committee)
 This was also an indication of an issue in the ERP implementation, which was fixed right away.

Details of Relevant Cases & Employees

Some of the Project’s Deliverables
Implementation Discoveries
►Unnecessary statuses
►Missing statuses & paths
►Missing business rules
►Lack of process control
Operational Benefits
►Visual deliverables assist in focused analysis
►Locating reworks and delays & possible bottlenecks
►Accurate throughput comparison

Mind Your Processes
Click, View & Do!

The Proposed Solution
Automatic process discovery (APD): looks at historical event log data and analyzes these data to generate visual models of an organization's business processes
Advanced business discovery technology: allows management to analyze & monitor operations and key performance indicators in real-time and alerts management of any anomalies

Intelligent Process Management & Visualization Technologies

Software AG’s Process Intelligence Solution
Software AG

Business Discovery with QlikView (BI)
The Business Discovery dashboard displays data & metrics based on sales data from the ERP system
The sales manager is not satisfied with the financial results in some countries
The manager wishes to analyze operational data in order to learn more about the way sales orders are handled.

The Process Model (in Priority ERP)

Process Discovery with QlikView
The Process Discovery sheet displays events & business data of closed sales orders from the ERP
Analysis of process performance by various dimensions is enabled
Gauges display operational info about duration, efficiency, costs and more
The current sheet is linked to the Business Discovery sheet.

Analysis of Slow Orders (> 40 days)
Upper bar chart displays sales orders durations on the X-axis, and number of orders on the Y-axis
Left-hand bar chart displays costs and durations per statuses or employees
Gauges view is updated according to user selections

Business Data of the Slow Orders
The Business Discovery sheet is automatically updated with the previous sheet selections
Only the slow orders-related data are displayed
Analysis of specific customers, part families or profit centers assists in discovering root causes of issues in the process

Further analysis of the slow process can be performed with Disco
Data of the slow process can be selected & exported with a click of a button.
Preparing Data for Automatic Process Discovery

The Data Exported from QlikView
The mandatory data for automatic process discovery

The Model of the Actual Process
The actual process flow discovered based on the imported process data
An Indication of inefficiency
Start point
End point

Animation of the Slow Process
Sales orders starting the process
An Indication of a Bottle Neck
An Indication of work loads
Event Time
The actual process is replayed based on the imported process data

Analyzing the Fast Orders for Benchmarking
The chart displays now data of sales orders with a duration of up to 15 days.
Comparing slow & fast processing of the sales orders can reveal opportunities for improvements (e.g., discover the efficient employees and adapt best practices)
The data is exported for further analysis with process mining technology.

Process Benchmarking
Fast Process
Slow Process

Real-Time Process Control
The Process Control dashboard displays both business & operational data of open sales orders
A deviation of case durations can trigger an alert (online or Email).

Business Rules Control
The dashboard enables to track deviations from business rules defined for the process
Gauges limits can be set dynamically
Alerts can be defined for deviations

Some of the Benefits
integrating automatic process discovery (APD) & BI technologies enable:
►Holistic approach vs. detached (i.e., a stand-alone solution/service)
►Event data are extracted directly from the BI database
►Event data are enriched with business data
►Analysis and benchmarking results enable fine-tuning of KPIs
►Operational & business managers can share results and insights
►Continuous process control & improvement

An Overview of the Solution
The solution enables the analysis of the performance on a production floor and it’s production process
It is possible to analyze overall production performance and individual machine performance across a number of key dimensions such as Work Center, Part, Time etc.
Automatic process discovery can be performed on selected data to identify root causes of deviations in costs, scrap, work orders cycle times and more
This demonstration is based on sample of production reporting and cost data of work cells, operations, products and machines from a metal processing company.

Manufacturing Analysis with QlikView*
* Based on the Plant Operations demo app

The Dashboard
The Dashboard enables analysis of the high level KPIs: Costs, Scrap and Runtime
Completed and Open work orders KPIs are displayed separately
 The data can be further analyzed by using the list boxes on the left of the sheet and the time dimensions at the top.

Analysis of Costs
The Cost sheet lets you analyze the costs of the plant & its operations.
With the charts in the container object, you can analyze Material, Overhead and Labor costs and Actual vs. Standard costs for various dimensions (Work Center, Operation and Part).
You can drill-down to focus on work order or operations with costs deviations

Analysis of Scrap
In the Scrap sheet scrap is displayed as a percent of production
Scrap can be analyzed over time and by key dimensions

Discovering the Actual Production Process
With Disco
Example Deliverables

The Planned Production Process

Some Possible Deliverables
A process map of the executed production processes
Conformance Checking
►Deviations from routings
Performance Analysis
►Machines
►Operators
►Idle times
►Bottlenecks
Focused analysis of:
►Breakdowns
►Rework
►Rejected parts
Comparison of workers performance

Export Production Data for Automatic Process Discovery
Automatic process discovery can be performed on selected data to identify root causes of deviations in costs, scrap, work orders cycle times and more
The data imported from QlikView with a click of a button.

Revealing Deviations from Routing
For some work orders, the process starts and ends with different operations then defined in the routing (circled in red)
Start point
End point

Locating the Slow Work Orders
The Performance filter is used to retrieve slow work orders by a selected duration

The Slow Process Flow
The mean waiting time between operations
The mean duration of operations

Idle Times in the Process
The numbers represent the maximal waiting time between work cells

Dafna Levy
Email: dafnal@nool.co.il
Phone: +972 (0)54-6881739
Intelligent Process Management
Thank You!

Intelligent Process Management & Visualization Technologies

More Related Content

What's hot (18)

Similar to Intelligent Process Management & Visualization Technologies (20)

More from Dafna Levy (20)

Recently uploaded (20)

Intelligent Process Management & Visualization Technologies