Upgrade to Pro — share decks privately, control downloads, hide ads and more …

QA Strategy in the AI Era

Avatar for Tomo Tomo
August 23, 2025

QA Strategy in the AI Era

QUALITY MEETS INTELLIGENCE: Evolving QA in the Age of AI🤖
09/08/2025
( https://www.meetup.com/arrangility-meetup-group/events/309146964/ )

#ARRANGILITY
#ArrangilityEvent
#TestAutomation
#AutomatedTesting
#QA
#AI

Avatar for Tomo

Tomo

August 23, 2025
Tweet

More Decks by Tomo

Other Decks in Technology

Transcript

  1. rrangility dn. hd.
 1
 QA STRATEGY IN THE AI ERA

    Robust test design centered on Value assurance Speaker: CEO Tomotaka Asagi
  2. © 2025 Arrangility Sdn. Bhd. AGENDA 2 1. Introduction 2.

    Conclusion New QA Strategy for the AI Era 3. Strategies detail 4. Human-AI Collaboration Model 5. AI-Accelerated Era QA: Action Guidelines
  3. © 2025 Arrangility Sdn. Bhd. Self-introduction Tomotaka Asagi CEO 3

    • Passion: Automated Testing • Hobbies: Photography, Driving • Work History: Sales → Development → Testing Software testing professional #ScrumMaster #CTAL-TM #CTAL-TA #CTFL
  4. © 2025 Arrangility Sdn. Bhd. Specializes in Automated Software Testing

    & QA Consulting 1. Introduction: My Company 6
  5. © 2025 Arrangility Sdn. Bhd. 1. Introduction: My Group Company

    7 Japan Quality Co., Ltd./ DAD Lab (Overseas verification base) 16F, Da Nang Software Park, 02 Quang Trung, Thach Thang Ward, Hai Chau District, Da Nang, VIETNAM ARRANGILITY SDN. BHD. Unit LG-02 & LG-03, Ground Floor, CoPlace 5 Block 3710, Persiaran APEC, Cyber 8, 63000 Cyberjaya, Selangor Darul Ehsan, MALAYSIA Human Crest Co., Ltd. 26F Yokohama Landmark Tower, 2-2-1, Minatomirai, Nishi-ku, Yokohama City, Kanagawa, JAPAN 220-8126 Japan・Vietnam・Malaysia
  6. © 2025 Arrangility Sdn. Bhd. 1. Introduction: Meet our team

    8 syazwanswong yuki-miyamoto19771210 fadzlianafarouk COO Yuki Miyamoto Sales Leader Solution Muhd Syazwan Solution Sales Nur Fadzliana Senior QA Engineer Mohamad Zaini mohamad z. Asian Regional Manager Aslina Aminuddin Aslina Aminuddin
  7. © 2025 Arrangility Sdn. Bhd. 2. onclusion - ew trategy

    for the ra 
 evelopment peed s ccelerating 
 any teams are experiencing dramatic improvements in delivery speed.
 
 uality ssurance s alling ehind 
 • esting struggles to keep up
 • requent test failures due to rapid changes
 • ifficulty verifying evolving requirements
 
 nconsistent egression esting 
 • arge disparities in automation maturity
 • esting frequency varies widely across teams
 
 nclear efinition of ervice alue 
 • any teams lack clearly defined quality expectations
 • ow clarity in what "value" truly means for end users
 10
  8. © 2025 Arrangility Sdn. Bhd. 2. onclusion - ew trategy

    for the ra 
 ssential trategies: 
 1. uality strategies tailored to the software lifecycle
 2. ystematize continuous requirement & value validation
 3. ormalization & optimization of regression testing
 4. obust test design resilient to change
 
 ey kill: 
 uality level articulation ability
 
 umans must clearly define quality standards to collaborate effectively with 
 11
  9. © 2025 Arrangility Sdn. Bhd. Quality strategy by software lifecycle


    1 Systematize Continuous Requirement & Value Verification 2 Normalize & Optimize Regression Testing 3 Design Robust Tests That Withstand Changes 4 14 STRATEGIES DETAIL
  10. © 2025 Arrangility Sdn. Bhd. Quality strategy by software lifecycle


    1 Systematize Continuous Requirement & Value Verification 2 Normalize & Optimize Regression Testing 3 Design Robust Tests That Withstand Changes 4 15 STRATEGIES DETAIL
  11. © 2025 Arrangility Sdn. Bhd. trategy 1 : uality strategy

    by software lifecycle 
 hort-term lifecycle 
 ✅ inimal quality assurance sufficient 
 ✅ ocus on verifying objective achievement 
 ✅ se -generated tests
 ❌ xcessive quality investment 
 16 ong-term lifecycle 
 ✅ ontinuous protection of value is essential 
 ✅ nvestment in robust test design 
 ✅ omprehensive quality assurance system 
 ❌ utting corners on quality 
 uman ecision riteria 
 xpected ifecycle × usiness mpact × ser xpectations 
 
 iverse oftware ifecycles in the ra 
 • ours to days: xperiments & prototypes
 • eeks to months: hort-term projects
 • ears and beyond: ong-term operational systems

  12. © 2025 Arrangility Sdn. Bhd. Quality strategy by software lifecycle


    1 Systematize Continuous Requirement & Value Verification 2 Normalize & Optimize Regression Testing 3 Design Robust Tests That Withstand Changes 4 17 STRATEGIES DETAIL
  13. © 2025 Arrangility Sdn. Bhd. trategy 2 : ystematize ontinuous

    equirement & alue erification 
 hy is it important? 
 • may misinterpret requirements 
 • alue may shift during development
 • isk of misalignment with stakeholders
 
 - pecific isks 
 • nintentional changes to service value
 • mplementation choices outside human intent
 18
  14. © 2025 Arrangility Sdn. Bhd. trategy 2 : ystematize ontinuous

    equirement & alue erification 
 19 mplementation pproach 
 • utomate requirement traceability
 • ontinuously monitor value indicators
 • egular stakeholder alignment
 
 ole ivision 
 • : utomate monitoring & detection
 • uman: et standards, judge, correct
 
 equirement alignment = rue quality 

  15. © 2025 Arrangility Sdn. Bhd. Quality strategy by software lifecycle


    1 Systematize Continuous Requirement & Value Verification 2 Normalize & Optimize Regression Testing 3 Design Robust Tests That Withstand Changes 4 20 STRATEGIES DETAIL
  16. © 2025 Arrangility Sdn. Bhd. trategy 3: ormalize & ptimize

    egression esting 
 21 hanges in the ra 
 • hanges occur daily or even hourly
 • npredictable side effects
 • ncreasingly complex dependencies
 
 imitations of raditional ethods 
 • eekly test execution
 • anual result verification
 • ifficult to predict change impact

  17. © 2025 Arrangility Sdn. Bhd. trategy 3: ormalize & ptimize

    egression esting 
 22 ecessary equirements 
 • ully automated (no human intervention)
 • ast execution (within minutes)
 • 24/7 continuous execution
 
 ptimization with 
 • uto-analysis of impact range
 • utomated test result evaluation
 • ynamic test selection
 
 
 heck overall integrity with every change 

  18. © 2025 Arrangility Sdn. Bhd. Quality strategy by software lifecycle


    1 Systematize Continuous Requirement & Value Verification 2 Normalize & Optimize Regression Testing 3 Design Robust Tests That Withstand Changes 4 23 STRATEGIES DETAIL
  19. © 2025 Arrangility Sdn. Bhd. trategy 4: esign obust ests

    hat ithstand hanges 
 24 imitations of raditional esting 
 • epend on implementation details
 • reak with changes
 • equire frequent maintenance
 
 mpact xamples 
 • 100% failure when is updated
 • nly 30% of business logic changes detected
 
 orsening in ra 
 • ncreased frequency of changes = increased instability

  20. © 2025 Arrangility Sdn. Bhd. trategy 4: esign obust ests

    hat ithstand hanges 
 25 - ra est esign 
 • entered around user value
 • irectly verifies core business logic
 • esilient to implementation changes
 
 mprovement ffects 
 • 0% impact from changes
 • 95% detection rate for business logic changes
 
 ests should not break when implementation changes 

  21. © 2025 Arrangility Sdn. Bhd. 3. uman- ollaboration odel 


    󰞴 uman ole: trategy & esicion 
 • efine lifecycle stages and required quality levels
 • larify business value and user value
 • stablish the test strategy
 • rticulate quality standards
 • onitor and decide value shifts
 27 🤖 ole: xecution & ptimization 
 • enerate tests based on defined standards
 • utomate execution and analyze results
 • earn patterns and suggest improvements
 • rive efficiency and optimization
 • ontinuously process large volumes of data

  22. © 2025 Arrangility Sdn. Bhd. 3. uman- ollaboration odel 


    󰞴 uman ole ( trategy & udgment)
 • ecide lifecycle and quality levels
 • efine business value & user value
 • ormulate test strategies
 • rticulate quality standards
 • onitor and judge value changes
 28 🤖 ole ( xecution & ptimization) 
 • enerate tests based on defined standards 
 • utomated execution & result analysis 
 • attern learning & improvement suggestions 
 • fficiency & optimization
 • ontinuous processing of large data 

  23. © 2025 Arrangility Sdn. Bhd. 29 hat oes ot hange:

    ev ps rocess 
 lan → ode → uild → est → elease → eploy → perate → onitor
 
 hat hanges: " uality tandard larity" 
 • ague requirements → → ague results
 • lear standards → → xpected quality
 
 rticulation ability determines quality 
 3. uman- ollaboration odel 

  24. © 2025 Arrangility Sdn. Bhd. 31 📱 hort-term (3-day ackathon)

    
 ecision: " asic functionality is sufficient"
 nvestment: 5% of development time
 esult: 100%
 
 🏢 ong-term (5-year ystem) 
 ecision: " ata integrity & are paramount"
 nvestment: 30% of development time
 esult: 300%
 
 🎯 uccess actor: lear judgment of required quality level 
 
 4. ractical xamples: ifecycle-based pproaches 

  25. © 2025 Arrangility Sdn. Bhd. 5. - ccelerated ra :

    ction uidelines 
 our lements of uccess 
 1. et appropriate quality levels based on lifecycle
 2. uild systems for continuous requirement/value validation
 3. ormalize and optimize regression testing
 4. esign robust tests resilient to change
 
 ctions to ake from omorrow 
 1. larify current project lifecycle
 2. et stakeholder agreement on expected quality levels
 3. ocument quality standards to communicate with 
 4. e-evaluate fragile tests from a value perspective
 
 33
  26. © 2025 Arrangility Sdn. Bhd. 5. - ccelerated ra :

    ction uidelines 
 he ngineer in the ew ra 
 rom a technical executor to a value-driven judge and communicator
 
 ngoing volution 
 • alance speed and quality
 • uild resilient validation systems
 • nsure continuous value delivery
 • ollaborate effectively with 
 34