internal audit data driven? · pdf file 2020-03-26 · vroege systemen waren...

Click here to load reader

Post on 08-Apr-2020

0 views

Category:

Documents

0 download

Embed Size (px)

TRANSCRIPT

  • Internal audit data driven?

    Edo Roos Lindgreen

  • abs.uva.nl

  • Programmes at the Amsterdam Business School  M.Sc.

     Executive Programme of Management Studies

     Executive M.Sc. in Finance and Control (RC)

     Executive M.Sc. in Internal Auditing (RO)  Executive M.Sc. in Insurance Studies

     AEMAS/APC (Actuariaat)

     Accountancy en Control in Deeltijd (ACD)

     M.Sc. in International Finance (MIF)

    Permanent Education Points

     MBA  MBA

     MBA in Big Data and Business Analytics

     MBA in Healthcare Management

     Postmaster  Postmaster Accountancy (RA)

     Executive Programme of Digital Auditing (RE)

     Executive Education  Lean Six Sigma

     The Analytics Academy

     Masterclasses

  • Executive M.Sc. Of Internal Auditing (EMIA)  2-jarige opleiding  72 EC  Theorie en praktijk  NVAO en SVRO accreditatie  Fasttrack voor RA’s  Docenten BKO-proof

     Auditing principles  Accounting information systems  IT basics, IT auditing, data science  Financial auditing, management

    accounting, internal governance  Managing the internal audit function,

    quality review, ethics and integrity  Audit skills, personal skills  Capita selecta  Internal auditing excellence

    Hoe ontwikkelt internal audit zich de komende 5 jaar?

  • Megatrends

    file://upload.wikimedia.org/wikipedia/commons/0/0d/Great_Wave_off_Kanagawa2.jpg

  • Top-10 risico’s voor top management

    Survey 1 1. Business interruption 2. Cyber incidents 3. Natural disasters 4. Market developments 5. Regulation 6. Fire and explosion 7. New technologies 8. Loss of reputation 9. Political risks and violence 10. Climate change

    Survey 2 1. Digital competition 2. Retaining top talent 3. Regulatory changes 4. Cyber threats 5. Resistance to change 6. New technologies 7. Privacy and security 8. Data analytics 9. Insufficient risk management 10. Customer loyalty

    CYBER

    DISRUPTION

    REPUTATION

    REGULATION

    TECHNOLOGY

    Allianz Risk Barometer 2018 Protiviti Executive Perspectives on Top Risks 2019

  • DSA1 1 - Introduction

    12

  • 1.000.000.000.000.000.000 100.000.000.000.000.000.000

    100.000.000.000.000.000.000.000

  • Cassandra Complex

  • Gartner Hype Cycle

  • RPA

    NOW LATER

    Data

    AI

    Algorithms RPA

    “Blockchain” Quantum

    IoT Bitcoin

    Process Mining

  • DATA

  • INTERNAL EXTERNAL

  • De data science epicyclus

     De goede vraag stellen  Exploratieve analyse  Predictive modeling  Interpretatie  Communicatie

    Peng. R. (2015). The art of data science.

  • Data-analyse en internal audit in de praktijk  Bestaand team “upskillen”  Team specialisten oprichten  Uitdagingen in samenwerking  In vijf stappen naar data-driven audit:

     Investeer in kennis en vaardigheden

     Investeer in mensen, infrastructuur en tools

     Stel de goede vraag; vind en ontsluit de data

     Maak de toegevoegde waarde van data-analyse zichtbaar

     Durf te experimenteren

    Data science for internal auditors

  • Data analytics tools en omgevingen

     Excel  BI tools

     Tableau, PowerBI, Spotfire  Algemene open-source

     R, Python, Jupyter, Knime  Proprietary

     SAS, Alteryx, SPSS

  • Algorithms, machine learning, AI

  •  Gebruik neemt toe (17%)  Convergentie RPA/AI  AI fte’s verdubbelen  Vereist goede organisatie  Governance  Control frameworks  AI as a service  AI onderscheidende factor

    Survey 1 Survey 2

     GDP +14% in 2030 door AI  Automatisering

     Productiviteitsverbetering

     Toename consumentenvraag

    Big 4 surveys

    KPMG. (2019). AI transforming the enterprise. PWC. (2019). AI: Sizing the prize.

  • AI en big tech  Big tech is ongeëvenaard

     Data

     Talent

     Geld

     Rekenkracht  Big tech companies stellen API’s

    beschikbaar: Google Tensorflow, Facebook PyTorch, Amazon AI (Lex, Polly, Rekognition)

  • AI eet meer rekenkracht op dan er bij komt

     Voor 2012  Gebruik van rekenkracht door AI

    verdubbelt elke 2 jaar

     Na 2012  Gebruik van rekenkracht door AI

    verdubbelt elke 3,4 maanden

     GPU’s

     Parallelle architecturen

  • AI rompsnelheid bereikt?  Vroege systemen waren gestructureerd, gebaseerd op modellen van mentale processen  Huidige deep learning systemen zijn ongestructureerde black boxes die worden getraind

    met enorme hoeveelheden data  Researchers Google, MIT: we bereiken de grenzen van deep learning, moeten een list

    verzinnen

    Battaglia, J.W. et. al. (2018). Relational inductive biases, deep learning, and graph networks. arXiv:1806.01261v3 [cs.LG] 17 Oct 2018. Hull speed = 4.49 x sqrt(length) in km/h

  • Internal audit�data driven? Slide Number 2 Slide Number 3 abs.uva.nl Slide Number 5 Slide Number 6 Slide Number 7 Programmes at the Amsterdam Business School Executive M.Sc. Of Internal Auditing (EMIA) Megatrends Top-10 risico’s voor top management Slide Number 12 Slide Number 13 Slide Number 14 Slide Number 15 Slide Number 16 Cassandra Complex Gartner Hype Cycle Slide Number 19 Slide Number 20 Slide Number 21 De data science epicyclus Data-analyse en internal audit in de praktijk Slide Number 24 Data analytics tools en omgevingen Slide Number 26 Slide Number 27 Slide Number 28 Slide Number 29 Slide Number 30 Big 4 surveys AI en big tech AI eet meer rekenkracht op dan er bij komt AI rompsnelheid bereikt? Slide Number 35