IBM Cloud Paks for Automation – Process Mining

Client Presentation

Stefano Pedrazzi
WW Process Mining Sales Leader
[email protected]

David Keyes
WW Automation Sales
[email protected]
IBM Cloud / © 2021 IBM Corporation
2
Intelligent automation
will improve business
performance by making in labor value will be released in 2022
all information-centric jobs by employing both automation and AI
more productive

of those scaling intelligent automation

expect their organization to outperform
the competition in revenue growth
You can’t
improve
what you
don’t know
Traditionally up to Only

of automation resources of automation projects

are spent pre-automation are able to scale up
Source: McKinsey, Changing Change Management Source: Deloitte - Annual Global RPA Survey Report
The intelligent automation journey

Cover all phases of the

intelligent automation
• Process mining
journey — don't jump • Task mining
to automation • Process
analytics

• Process
modeling
• Process
documentation

• DPA
• RPA
• Low code

Source: Create a Governance Strategy To Meet The Process Mining Imperative

Process mining as your navigation system
IBM Cloud / © 2021 IBM Corporation
7
Process mining—the real processes derived from data

The customer view of the

organization and processes

Sources

Mobile IBM Process

Mining

Desktop

Data Knowledge
storage base

Cloud
Improve business processes starting
from the digital footprint

Data: Logs from information systems Data: Digital twin

User’s desktop
recordings

Scope:
Actual work done
by employees
Analytics
Extract Task mining Automate

Event logs (Data sources) Process Discovery

mining
IBM Cloud Paks for Automation
The most complete set of business and IT Automation services, powered by AI

Use cases App and infrastructure stability Hybrid cloud integration Site deployment automation
Customer experience

Business operations Platform operations Real-time interactions 4G/5G telco cloud platform

Workforce management AIOps optimization Transactional integrity NFV lifecycle management

IBM Cloud IBM Cloud Pak for IBM Cloud Pak for IBM Cloud Pak IBM Cloud Pak for
Paks Business Automation Watson AIOps for Integration Network Automation
⎻ Workflow and decisions ⎻ Application impact avoidance ⎻ Application integration ⎻ Intent-driven orchestration

⎻ Content services ⎻ Hybrid application management ⎻ API management ⎻ Closed-loop operations

⎻ Operational intelligence ⎻ Observability ⎻ Messaging and events ⎻ Network optimization

Automation foundation

⎻ Robotic Process ⎻ Natural language ⎻ Process and task ⎻ Event detection ⎻ Machine learning ⎻ 3rd Party ⎻ Operational models
Automation interactions mining integrations

IBM Cloud AWS Microsoft Google VMware Private IBM Z End

Azure Cloud IBM LinuxOne points
IBM Power
Systems
From process mining to a winning automation plan

Use the process insights to

design your automation strategy

– Identify repetitive and Automation Actions

inefficient tasks in need
of automation Examples of insights
Robotic Process Automation
– Analyze process Automate high-frequency manual activities, reworks
performance, productivity,
and frequency challenges Automation Decisions

– Simulate process Reduce non-conformant decisions

enhancements, prior
Workflow
to implementation
of automation Reduce process deviations and variants

Document Processing

Eliminate manual document processing bottlenecks

Trigger Actions on Apps

Predict late running cases

AI Ingestion
Why is IBM different?
IBM Process Mining can derive an extremely precise model of your
processes and enable you to pursue continuous improvement initiatives
with a low-cost and fail-fast approach.

KEY DIFFERENTIATORS

Digital twin Decision Simulation Process mining

rules miner and task mining
IBM Process Mining Play with the Digital Twin
visualizes complex IBM Process Mining of an Organization, create A seamless integration that
processes in a simple way. discovers not only the what-if scenarios of the accelerates and scales the
different routes of your to-be process, and adoption of automation,
Most advanced and
process, but why the estimate the expected ROI combining business
precise model discovery
process is following of automation initiatives. activity analysis with
on the market.
a route. task analysis.
IBM Process Mining
named a Leader in the
NelsonHall Insight's 2021
NEAT Report for Process
Discovery & Mining
Process mining
Process mining
From real events logged by enterprise apps

– Discover processes from real work

Intelligent Automation from facts
– Discover lead time, cost, rework,
Shipped with Cloud Pak for Automation foundation and conformance issues
– Discover automation opportunities

Available on Connection Task mining

– Red Hat OpenShift points examples From human actions
(operator) – Workflow interoperability
– Discover real work done during an activity
– On-premises – Business Rules
– Automate most frequent action paths with RPA
(traditional installer)
– RPA recommendations
– Combine process and task mining to analyze
– Software as a Service
– Data integration different levels of details, and get a complete
– AI ingestion picture of your process
Process mining journey
The process 1 2 3

Process setup Process mining Task mining

mining journey Scope of the analysis, Events log User actions recording
data extraction – Frequency view, performance view, costs – Automated analysis of steps executed by
– Path and deviations analysis users to perform process activities (tasks)

– Lead time influencers, SNA, activity map – RPA candidates

4 5 6

Compliance Automation KPI and

Define Reference Model (BPMN) analysis performance
– Violations and root cause analysis Define standard costs Define KPIs
– Segregation of duties – As-is automation and candidate tasks – Verify standard/custom KPI’s compliance
for automation – Process compare and best practices

7 8 9

Simulation Real-time and Automation plan

Scenarios and goals definition predictive analytics – Act on insights
– Allocate RPA bots
– Simulate what-if scenarios – Real time monitoring of running cases
– Achieve end-to-end process automation
– Compare as-is and to-be – Estimate lead time and costs
– Measure automation performance
– Compute automation ROI against business goals
Data extraction and process setup

IBM DataStage,
ERP
App Connect,
Talend, …
CRM Data source, data mapping
Extract
Transform
Transactions HR
Load (ETL)*

BAW

BAI IBM Process

Mining

Task Mining
User desktop
actions

*ETL contribution packages available for P2P/SAP and O2C/SAP

Discover
frequencies,
variants, rework,
durations, costs
from real data

– Instantly visualize
the real processes

– Most frequent activities,

reworks, and variants
can be good automation
candidates for RPA
or Workflow

– Highest wait-time
and durations show
potential bottlenecks

– Identify costly activities

based on resource cost
Check
conformance
with reference
models
– Quickly identify
non-conformant cases

– Import as-designed BPMN

reference models from
any Business Process
Automation Tool that
supports BPMN

– Discover cost of
non-conformance,
fitness level Non-conformant activity
– Run a root-cause analysis and transitions

List of deviations and statistics

Generate
BPMN from
real processes
– BPMN model generated
from events, filters can
be applied (exclude
running cases, exclude
non-conformant cases,
keep most frequent
variant, etc.)

– Export BPMN to
any Business Process
Automation tool that
supports BPMN for
finer modelling

– Export BPMN to Business Process Process Process

Process Management tool Roles Activity Gateway
that supports BPMN for
implementing workflows
Generate DMN
decision tables
from real
processes with Statistics of
rules mining coverage and
precision of
– Rules mining discovers
each business
the busines logic behind rule discovered
the gateways

– Set the business data to

consider the details level

– Implement the business

logic in any Decision
Management software

Business rule discovered

Set and
monitor KPIs

– Set KPIs based on case

lead-time and cost, activity
throughput, activity wait-
time, resource allocation
and cost

– Monitor in process views

and dashboards Critical Activities
Monitor
processes with
dashboards
– Generic overview
dashboards

– Customizable business
dashboards built from
widgets with low-code
user interface

– Out-of-the box Procure to

Pay and Order to Cash
dashboards such as
deviations, maverick
purchases, cash discount 30+ widgets like
losses and cash flow losses
bar chart, bubble
chart, line chart
to be customized
Simulate your
business with
digital twin
before engaging
development
costs
Compare the simulation with
– Create scenarios from
as-is process: allocating the data derived analysis
resources, adding RPA
bots, changing processes,
and rules

– Run simulations and

compare as-is and to-be
side-by-side: processes,
KPIs, costs

– Load BPMN models from

from any tool that supports
BPMN and run simulations
to verify achievement of Select the expected
business goals level of automation of
a specific activity
Task mining – How it works How does task
mining work?

Capabilities: Users Needs

– Monitoring list – Events analysis
and correlation Record user
– Context recognition (OCR) interactions
– Automatically transmitted
– Advanced custom to process mining for
recording use by analyst

Context
Benefits of task mining and recognition
process mining integration:

– Perform analyses at different levels of detail for

your process, providing you with a complete picture
of your process Exploit task
mining potential
– Use the insights to design the best automation strategy.
Automate frequent paths of the process, and repetitive
and unproductive tasks
Upgrade
– Design strategies to reduce the most critical process your Digital Twin
of an Organization
deviations by focusing on the related root causes
Task mining: Record user tasks to drill-down into process
activities or when application logs not available
a use case 2 3

1. Process mining tells us

the service time on the 1
“BO Service Closure”
activity. How can we
discover the tasks that the
users usually execute to
complete this activity?

2. Task Mining records the

user interaction and 4
displays the tasks as a
sub-process of the BO
Service Closure activity.

3. For each task, we know

exactly the productive time,
idle time, and time spent
on other tasks.

4. By combining Process
Mining and Task Mining
data we can get automation
recommendations.
Process mining
use cases
Case study: A digital transformation journey

Italian Bank: Credito Emiliano (Credem)

Process mining is used together with RPA, workflow and

ODM to manage the bank’s overall digital transformation
journey including process improvement and
automation initiatives

Challenges: Results:
– Improve internal efficiency
– Improve customer experience 85% Cost savings in back-
office service closures

50+ 400+ 30% Process lead

time reduction

Rolled out
processes
Targeted
processes 20% Overall cost
reduction
Case study: After sales customer service

Fashion company:

Challenges:
– Reduce time
– Cost of customer request management

Solution: Results:
– Process discovery and analysis
– Request classification with AI
– Solution automation with RPA
90% (27 days) Completion
time average reduction

46% Average cost

saved per case
Case study: Procure to pay in manufacturing industry Results:

Unleash trapped value

in your P2P process
47%
Reduction of average
invoice registration time
47% 1M$

Area of impact
9%
FTE savings for
Use case result
Automation
Working country of Italy
capital

$1M
Global increase of
Approve Goods and cash flow availability
Purchase Purchase
request order
and send services Invoice Payment for every $1B spent
PO receipt

9% Maverick Discount
Deviations
buyings losses
Next steps
IBM Process Mining

From problem finder to a self-driving organization

Insight Action Resilience Self-driving

Process intelligence Process execution Process adaptability Process independence

– Automated process – Leverage the IBM Cloud – Full DTO to consider – Autonomous
discovery and mining Paks for Automation organization dependencies hyperconnected processes
to deliver best of breed
– Process performances, AI-powered automation – Automated impact analysis – Self-automation
automation, and costs and improvements
– Common low-code, – Automated alignment
– Compliance/conformance no-code, and of skills and automations
– Root/cause analysis orchestration platform

– Simulation, what-if – Explainable AI to

and ROI forecast justify decisions

Today Future