AI in Test Automation

®
WEBINAR: Automation Demystified
Part 7: AI in Test Automation. When to expect
The Raise of Machines?
• August 7th, 2019 – Alexey Grinevich, alexey@inflectra.com
We will be starting the webinar shortly, please stand by…
All phones will be automatically on mute until the Q&A.

®
2 | 8/7/2019 © Copyright 2006-2019 Inflectra Corporation
Webinar Series: Useful Information
• Housekeeping Matters
• Post Webinar Follow-Up:
• Blog
• Webinar Recording
• Presentation On Slideshare
• Post-Webinar Quiz
• Webinar Series: Test Automation,
Demystified
• Webinar Series Certificate Rules &
FAQs

®
Hey, I am Alexey
• Working on Rapise test
automation tool @Inflectra
• 20 years in software
development
• 10 years in test automation
• Twitter: @AlexeyGrinevich
• Email: alexey@Inflectra.com

®
Topics
 Levels of Testing Automation
 Cars vs Software
 What is Testing?
 Swim Lanes
 Current Level
 When to Expect Level 4?
 What to Do About It?

®

®
Same for Automated Testing?
Driving Software Testing
A Car Software Application
A Driver Human or Robot Tester
The Road Network ???
From Address 1 to Address 2 Perform Testing

®
The Road

®
The Road in Software Testing
 Lanes are Forms, Controls and API entry points
 Road signs and Map are specifications and documentation
 Route is a scenario

®
The Goal
Get from Address 1 to Address 2
 
Perform Testing

®
What is Testing
 Finding bugs.
 Looking for crashes.
 Checking that application can be used for its purpose.
 Checking that application confirms to the specification.
 Deciding if an application may go to the production.
 Validating that application is secure and does not expose private or restricted information
outside (security testing).
 Checking that application is able to serve required load (load testing).
 Comparison testing (to similar product).
 Regression testing (compare to previous version of the same product)
 Localization testing (product works for multiple countries).
 Usability testing.
 Documentation testing.
 … more …, much more.

®
So Testing Is…
 No Common Meaning
 One of the goals – define testing for specific application
 Currently Done Manually

®
4 Swim Lanes
 Software Interaction
 Scenario Creation
 Model / Specification Creation
 Goals / Targets / Acceptance Creteria

®
Swim Lane State
No AI (Manual)
Partial AI
Full AI

®
Swim Lane 1: Software Interaction
 Find controls on the screen.
 Click / Select / Set Text / ETC
 Most well developed lane.
 Partial AI and expect Full AI

®
Swim Lane 2: Scenario Creation
 Create Scenarios based on
 Specifications
 Documentation
 Logs / Traces
 Monkey Testing
 Currently No AI to Partial AI

®
Swim Lane 3: Model / Specification
 Formal Specifications
 Gherkin/Cucumber
 Diagrams
 Finite State Machines
 Currently No AI to Partial AI

®
Swim Lane 4: Goals and Criteria
 Sometimes goals are known
 100% Code Coverage
 99.9% Fault Tolerance
 Size Limitations
 defined by Tester or QA eng.
 Currently No AI

®
Summary on Automation
Interaction Scenarios Model / Spec Goals / Targets
Manual Level 1
Partial Level 2
Full Level 3 Level 4

®
What is Currently Done
Rapise Applitools Functionize Mabl ReTest Test.ai TestCraft Testim
From Natural Language V V
Find elements like a Person V
Visual Testing V V V
Self-Healing Locators V V V V V V V
Monkey Testing V

®
Where are we?
 Level 2
 Some Swim Lanes moving to Level 3.

®
Rapise
 Self-Healing
 Moz 500 Analysis
 Web Profiles

®

®
Moz 500

®
Level 4: When?
 Level 3: When?
 Let’s look into history

®
Trains
 London Underground's Victoria line – 1967
 ATO (Automatic Train Operation)
 Grade of Automation (GoA) GoA 0 to GoA 4
 There are several lines operating at GoA 4 (unattended train
operation) and more to appear within next few years

®
Auto Pilot
 Was there in 1930
 But still no fully automatic
passenger air travels

®
Infrastructure
 Road quality, regulations
 Supplementary technologies (sensors, scanners, EBS, ABS,
Cruise Control, ML, Computer Vision)
 Global Infrastructure:
 GPS / Glonass / Galileo
 CAT IIIb landing or Autoland
 Demand. The way we use cars changes
 Bikes
 Public transportation
 Uber
 High speed trains

®
How it Affects Software Automation?
 Infrastructure actively changes
 Quick Delivery
 Test by Users
 Software is Flexible
 Microservice Architecture
 Lass fault tolerant
 Flaky by Nature
 Accessibility Standards
 Iaccessible
 ARIA

®
What Affects AI Adoption
 Infrastructure
 Short Release Cycles
 AI
 So Primary Focus is:
 Application Lifetime Management (Spira)
 Automated Test Maintainability (Rapise)

®
Summary
 So When Is Level 4?
 Sir Antony Hoare
 Null pointer
 Quick Sort
 Hoar monitors

®
Global Challenge
 The Verifying Compiler: A Grand Challenge for Computing
Research’
As an example drawn from Computer Science, I revive
an old challenge: the construction and application of a
verifying compiler that guarantees correctness of a program
before running it.
 Way to Avoid Programmers and thus Testers
 2055

®
The World Changes an Unexpected Way

®
Stay Calm
 Machine Learning vs AI
 Boston Dynamics (no ML)
 AlphaGo (19x19 -> 9x9)

®
Questions are Welcome
Contact me
• Twitter: @AlexeyGrinevich
• Email: alexey@inflectra.com
• Web: https://www.inflectra.com/Rapise/

®
Q&A
Thank You for Watching
Subscribe to Our Channel for More Videos on Software Testing

AI in Test Automation

More Related Content

What's hot (20)

Similar to AI in Test Automation (20)

More from Inflectra (20)

Recently uploaded (20)

AI in Test Automation

Editor's Notes