Upload
others
View
0
Download
0
Embed Size (px)
Citation preview
us.sogeti.com
COGNITIVE QA: LEVERAGE AI AND ANALYTICS FOR GREATER SPEED AND QUALITY
COGNITIVE QA: LEVERAGE AI AND ANALYTICS FOR GREATER SPEED AND QUALITY
us.sogeti.com
ARTIFICIAL INTELLIGENCE vs. COGNITIVE COMPUTING
2
Build a system that
can generally perform
any intellectual task –
so called Strong AI
Build a system that
perform can an
intellectual task in
smaller domain – so
called Weak AI
Ability to learn and build
models, so that a system
perform activities like
prediction within specific
domains
Build a system that can
learn and interact in
natural way with humans
Emerging of Deep
Learning with neural
networks with many
layers, new topologies
and learning methods
Artificial Intelligence
Machine Learning
Cognitive Computing
1950 1960 1970 1980 1990 2000 2010
us.sogeti.com
Structured and active Unstructured and dark
Data that’s coming
• Customer records• Transactional systems
• Predictive models
• Institutional expertise• Operational systems
• News• Events
• Social media
• Weather• Geospatial information
• Internet of Things (IoT)• Sensory data
• Images
• Video
Data outside your firewallData you possess ++
CONNECTING DISPARATES TYPES OF DATA WITHIN AND OUTSIDE YOUR WALLS CREATES NEW OPPORTUNITIES FOR UNEXPECTED INSIGHTS
us.sogeti.com
WITH WATSON YOU CAN HARNESS THE POWER OF AUGMENTED INTELLIGENCE
Your data + the world’s data
Watson can ingest, cleanse, and comprehend multiple data types. Build off your domain expertise and the largest base of industry offerings in the market.
Ingests + transforms
Apply a powerful suite of Watson services and APIs to transform your data into valuable insights. Watson is optimized for robust AI workloads and lives on IBM Cloud.
Insights + outcomes
Watson extracts meaning to provide deep insights and produces powerful outcomes. Train your AI on what’s important to your industry gain knowledge and make the most informed, reasoned decisions.
StructuredUnstructured
Transform
Insights
us.sogeti.com
IBM WATSON IS THE AL PLATFORM FOR BUSINESS
Watson is the AI platform for business. With Watson
you can discover hidden insights, engage in new ways, and decide
with greater confidence.
us.sogeti.com
BY APPLYING COGNITVE CAPABILITIES TO YOUR DATA, YOU WILL ENHANCE DIGITAL INTELLIGENCE EXPONENTIALLY
REASON
They can reason, grasp underlying
concepts, form
hypotheses, and infer and extract ideas.
UNDERSTAND
Cognitive systems understand
imagery, language
and other unstructured data
like humans do.
LEARN
With each datapoint, interaction
and outcome, they
develop and sharpen their expertise, so they
never stop learning.
INTERACT
With abilities to see, talk and hear,
cognitive systems
interact with humans in a natural way.
us.sogeti.com7
WHAT CUSTOMERS ARE SAYING
“Intelligent test automation and smart analytics will become essential
to support testing as they enable smart decision making, fast
validation and automatic adaptation of test suites.”
World Quality Report 2017-18 findings suggest
How to ensure sufficient
quality without slowing down
on speed?
Which test cases are
relevant and which ones
are not?
How to improve testing with
user behavior patterns?
How to keep up with increasing complexity of
intelligent applications?
us.sogeti.com
KEY QA CHALLENGES TODAY
• Validating the increasing complexity of the application landscape is becoming more and more difficult
• Keeping pace with the demand for increasing speed to market is challenging
• Test sets are either missing up to 30% of relevant tests and/or containing many irrelevant test scenarios
Increasing speed and complexity
• Automation is currently under-exploited in QA & Testing and as a result there is too much reliance on manual activities
• Traditional automation approaches have delivered only limited benefits
• Smart applications demand intelligent automation, integrated in continuous delivery
Insufficient Automation
• Test environments and test data are the most cited challenges for core enterprise applications as well as more fluid IT applications
• Smart test platforms are required to deal with the growing challenges of test environments, data and virtualization
Test environments and test data
• Usage patterns and user feedback are not exploited for QA purposes
• Customer experience validation is a challenge
• Social media and other external data sources are seldom leveraged for QA purposes
Client focus inside out
us.sogeti.com9
COGNITIVE QA VALUE
• Leveraging self-learning and analytical technologies for
• Predictive QA Dashboards
• Smart Analytics for QA
• Intelligent QA Automation and
• Cognitive QA Platforms
• Enabling smart quality decision making on factual project data, actual usage patterns and user feedbacks
• Delivering quality with speed in a complex connected world at optimized cost.
COGNITIVE QA
INTELLIGENT QA
AUTOMATION
COGNITIVE QA
PLATFORMS
PREDICTIVE QA
DASHBOARDS
SMART ANALYTICS
FOR QA
With Cognitive QA organizations will be able to achieve
an accelerated and optimized quality by using an
intelligent approach to QA>
us.sogeti.com1010
DEVELOP A VISION AND ROADMAP
Predictive QA Dashboards
• X-application transparency
• Near real-time insight
• What-if prediction
Smart Analyticsfor QA
• Efficiency through focused Testing
• Testing aligned with actual usage patterns
• Improved test coverage
• Improved risk management
Intelligent QA Automation
• Maximized RoI on test automation
• Focused test automation aligned with actual usage patterns
• Accelerated, touch-less testing.
Cognitive QA Platforms
• Instant validation.
• Optimized test coverage with self adapting test suites.
• Minimized idle time and cost with self-aware and adaptive environment provisioning.
• Optimized risk management with self-learning capabilities.
Accelerated, intelligent QA
leveraging AI and Analytics.• Modular.• Innovative.• Business oriented.
Future Proof
QA
Guidance to the journey towards
quality with speed.
us.sogeti.com
PREDICTIVE QA DASHBOARDS
11
Leverage all relevant data from Dev / Test / Ops repositories for:• 360o Visibility – get comprehensive and transparent views on application quality on management and operational level.
• Tailor-made Insights – customized for your specific demands to allow better control and steering of quality across all your projects.
• Near Real-Time Updates - refreshed continuously to enable timely and improved decision making.
• Zoom In – drill down in details to do root cause analysis of real issues and analysis on expected outcomes.
• Alerts – avoid surprises with pro-active notifications to minimize the risk of serious errors.
• What-if Analysis – optimize test strategy in line with available resources and coverage requirements.
• Predictive Analysis – leverage historical usage patterns by applying proven algorithms for defect prediction and test coverage optimization.
us.sogeti.com
Implemented step-by-step to solve your Business Challenges
•Automated Test Case Selection and Prioritization – gain efficiencies through focused testing effort.
• Predictive Test Environment Configuration – minimize idle time with optimized environment provisioning.
• Test Automation Prioritization – maximize ROI of test automation effort.
• Test Coverage Optimization – make best possible choices based on ‘what if’ analysis.
• Test Execution Analysis – align test with actual usage patterns.
• Auto Generation of Test Scripts – create optimized test suites while reducing preparation effort.
SMART ANALYTICS FOR QA
12
Proven Analytics Solution Elements delivering Real Business Benefits
> 40 tried and tested rules
• Rules of Thumb
• Common Sense
• Best Practices
• Intuitive Judgments
• Positive / Negative Patterns
Applying our IP
Kick-StartKick-Start Proof of ValueProof of Value DeploymentDeployment
• Collecting basic information using Cognitive QA Pre-Assessment questionnaire.
• Checking availability, accessibility and quality of data.
• Performing Solution Build Workshop to define the scope of POV.
• Collecting basic information using Cognitive QA Pre-Assessment questionnaire.
• Checking availability, accessibility and quality of data.
• Performing Solution Build Workshop to define the scope of POV.
Conducting Proof of Value with following steps:• Verify analytics readiness.• Data collection and preparation.• Modeling Cognitive QA rules.• Validation and reporting.• Define the deployment approach.
Conducting Proof of Value with following steps:• Verify analytics readiness.• Data collection and preparation.• Modeling Cognitive QA rules.• Validation and reporting.• Define the deployment approach.
Operationalizing rules implemented during PoV. Major activities involved:• Implement connectors to pull live
data.• Implement orchestration engine to
automate workflows.• Implement portal (responsive web).
Operationalizing rules implemented during PoV. Major activities involved:• Implement connectors to pull live
data.• Implement orchestration engine to
automate workflows.• Implement portal (responsive web).
us.sogeti.com
Source ctrl /Build Mgmt System
Project Mgmt System
GrouperGrouper
Data RulesData Rules
Software Lifecycle Artifact
Quality
Software Lifecycle Artifact
Quality
Non-structured
Data
structured
Data
Test & Dev
Repositories
Self Learning
Recommendation Engine
Self Learning
Recommendation Engine
TestOps
Rules
TestOps
Rules
Test Results
Optimized Test list based
on real time data
Test Engine
Feedback
Test Mgmt System
Req. Mgmt system
Code Quality Parameters
Operations
Repositories
RTM Systems
Usage patterns
Support engr call notes
App. Logs and traces
structured
Data
Non-structured
Data Defect Mgmt System
SOLUTION CONCEPT OF SMART ANALYTICS
us.sogeti.com
TOOLS AND TECHNOLOGIES
14
• Automatic Regression Test Prioritization and Execution
‘PoV in a box’-assets• Pre-assessment questionnaire• Solution Build workshop template• Assessment checklists• Cognitive QA rules
• Report template• Customizable Dashboard with Standard Metrics• Proof of Value references
IN-HOUSE ASSETS
INDUSTRY TECHNOLOGIES
us.sogeti.com
CASE STUDY
15
INDUSTRIAL AUTOMATION OEM CUSTOMER
• Dynamic Regression Test Selection • PoV duration 10 weeks• 6 Cognitive QA rules implemented • The Total number of Test Cases were 22,289 ,
and regression identified for the current release was 8,773.
• We recommended 2,288 regression Test Cases.
• Dynamic Regression Test Selection • PoV duration 8 weeks • 6 Cognitive QA rules implemented• The Total number of Test Cases were 13, 721 • We recommended regression Test Cases of
1, 664.
• Selection of Configurations and Tests that matter
• PoV duration 10 weeks• 8 Cognitive QA rules implemented• We recommended: 325 high priority & 571
low priority test suites of 896 total test suites.
• We recommended : 44 high priority & 141 low priority configurations among 192 Configurations.
• ~73% effort saving, shorter release cycle
• Team gets more time to automate
• The high risk test cases are executed in the
beginning of the test cycle, early detection of
regression defects.
• Fixing the identified tool and process gaps will
prepare for better “Analytics Readiness”.
• ~73% effort saving, shorter release cycle
• Team gets more time to automate
• The high risk test cases are executed in the
beginning of the test cycle, early detection of
regression defects.
• Fixing the identified tool and process gaps will
prepare for better “Analytics Readiness”.
KEY BENEFITS
• Improved Adaptive Regression Test Selection
strategy factoring multiple inputs.
• More quality builds for special purpose testing
like system testing, performance testing.
• Recommended test list is being automated and
made part of regular automated test
executions.
• Fixing the identified tool and process gaps will
prepare for better “Analytics Readiness”.
• Improved Adaptive Regression Test Selection
strategy factoring multiple inputs.
• More quality builds for special purpose testing
like system testing, performance testing.
• Recommended test list is being automated and
made part of regular automated test
executions.
• Fixing the identified tool and process gaps will
prepare for better “Analytics Readiness”.
• Helped identify 63% test cases that can be
deprioritized.
• Derived critical tests and configurations with
high probability of risk.
• Systematic way of System under Test
configuration planning .
• Better utilization of the hardware/ Cost etc.
• Helped identify 63% test cases that can be
deprioritized.
• Derived critical tests and configurations with
high probability of risk.
• Systematic way of System under Test
configuration planning .
• Better utilization of the hardware/ Cost etc.
KEY BENEFITS
MEDICAL IMAGING SOLUTION PROVIDER
KEY BENEFITS
us.sogeti.com16
THANK YOU
us.sogeti.com17
BACKUP
us.sogeti.com
THERE ARE LOTS OF USE CASES
Improve coverage!
Optimize and align test sets based on real application usage
Test Coverage Optimization
Release Readiness
Predict and advice marginal value of additional testing
Automated impact Analysis
Quality Dashboard
Cost of a bug, Test efficiency, Release Dashboard
Predictive quality metrics
Auto-Heal
Throw away the Bad/ obsolete test cases (future proof testware )
Isolate Script Failures
Throw away the bad code fixes in CI pipeline
Automated Test Selection & Prioritization
Predictive test environment configuration
What should we automate?
Tests that matter!
Resource utilization
Auto Generation of Test Scripts
Predict risk level of product components / services / systems (incl. UX)
Predictive capacity planning
us.sogeti.com
RULES SELECTION