Slide 1: Appen Limited Annual General Meeting
Appen Limited Annual General Meeting 31 May 2019
Slide 2: Forward looking statements
The forward looking statements included in these materials involve subjective judgement and analysis and are subject to significant uncertainties, risks, contingencies, many of which are outside the control of, and are unknown to Appen Limited. In particular, they speak only as of the date of these
Slide 3: Chairman's welcome
Chairman’s welcome eappEen| 3
Slide 4: Agenda
Agenda • Chairman’s address • CEO presentation • Address from Nomination & Remuneration Chair • Formal business and resolutions • Morning tea 4
Slide 5: Board of directors
Board of directors • Mr Chris Vonwiller Chairman • Mr Bill Pulver Non-Executive Director • Ms Robin Low Non-Executive Director • Ms Deena Shiff Non-Executive Director • Mr Mark Brayan Chief Executive Officer Apology: • Mr Steve Hasker Non-Executive Director 5
Slide 6: Chairman's address
Chairman’s address #aPpPEN'| «
Slide 7: CEO's presentation
CEO’s presentation 2appEen| 7
Slide 8: FY2018 highlights (A$M)
FY2018 highlights (A$M) Revenue Appen’s strong growth continues 364.3 • Revenue up 119% to $364.3M • Underlying EBITDA of $71.3M up 153%, statutory EBITDA up 206%1 166.6 • Underlying EBITDA margins improved from 16.9% to 19.6% • Underlying NPAT of $49.0M up 148%, statutory NPAT of $41.7M up 192%2 FY
Slide 9: Strong divisional performance (A$M)
Strong divisional performance (A$M) Substantial 2H for Language Resource High revenue growth and margin delivers record annual revenue expansion for Content Relevance Language Resources Content Relevance 51.4 312.8 40.4 Revenue Revenue EBITDA 126.2 EBITDA 12.2 11.2 74.8 22.1 FY2017 FY2018 FY2017 FY2
Slide 10: High growth, long term customer relationships (A$M)
High growth, long term customer relationships (A$M) 364.3 Ongoing customer growth and repeat revenue • New and existing customers growing with multiple projects • Consistent growth underscores importance of 166.6 data and quality of Appen’s delivery 110.9 • Growth driven by data for new projects, mo
Slide 11: Business initiative updates
Business initiative updates Secure facilities Leapforce • Growing demand due to data privacy and • Integration virtually complete commercial confidentiality • Increased automation and scalability for all • Manila facility ISO 27001 certified operational processes under Appen Connect (prior Leapforce
Slide 12: Additional growth initiatives
Additional growth initiatives Customer-centric alignment China • LR and CR sales teams combined to unify • Largest AI market outside US customer engagement and support • Building out team, customer base and operations • Operations teams combined under global in Shanghai Client Services team to impro
Slide 13: Transformative acquisition
Transformative acquisition Leading global crowd Best in class SaaS platform Integrated global data annotator • Over 1m on-demand global crowd • Unique ‘human-in-the-loop’ • Vertically integrated combination platform capabilities transforming with global crowd solutions utilising • Covers 130 countri
Slide 14: AI mimics human behaviour, typically two approaches
AI mimics human behaviour, typically two approaches Humans provide the rules Humans provide examples We support leading AI applications by providing ‘examples’ (aka training data) 14
Slide 15: To learn, computers are 'told' what data represents
To learn, computers are ‘told’ what data represents This sequence of numbers is an image of Abraham Lincoln 15
Slide 16: More training data = better performance
More training data = better performance Source: Andrew Ng 16
Slide 17: Training data is the 'product' when it comes to AI
Training data is the ‘product’ when it comes to AI Design Code Test Iterate Improve the Start with logic performance by (e.g., Abraham Take the logic and Use real world Old approach Lincoln has a beard convert into code examples to test adding/changing logic and corresponding but no moustache) code
Slide 18: Appen provides high-quality AI training data
Appen provides high-quality AI training data Data Collection Machine learning models require large volumes of high-quality data to be trained effectively. Scale your data collection efforts across multiple file formats including text, image, video and speech. Data Annotation Annotated data enables r
Slide 19: We focus on three main data types
We focus on three main data types Relevance Speech & Natural Language Image & Video • Enhance on-site search, • Data for speech recognition, • Data collection and annotation categorization, personalization speech synthesis and natural for computer vision and more language understanding • Use cases i
Slide 20: We deliver at scale
We deliver at scale Scale >1m crowd, >10 billion data points to date Diversity >180 languages, >130 countries, >70k locations Social impact Access to ‘Differently Abled’ workforce 20
Slide 21: Leading technology to enable our crowd and customers
Leading technology to enable our crowd and customers Global crowd workforce Crowd Client Annotation management workspace tools • Recruiting • Collect and load data • Speech and text • Onboarding • Design job • Relevance • Quality assurance • Select workers • Image and video • Payments • Launch and m
Slide 22: Appen is strongly positioned in a high growth market
Appen is strongly positioned in a high growth market Established Provider Global Presence Scalable Operations Founded in 1996 with industry-leading work in 600+ full-time staff speech solutions worldwide Added search relevance Exeter High staff engagement, Seattle Beijing expertise in 2007, Detroit
Slide 23: Outlook
Outlook Appen is strongly positioned and continues to execute in a high growth market Appen’s core business is performing well and Figure Eight is meeting expectations YTD revenue plus orders in hand including Figure Eight for delivery in 2019 ~$270M at mid-May 2019 The Company’s full year underlyin
Slide 24: Thank you
Thank you Mark Brayan, CEO mbrayan@appen.com Kevin Levine, CFO klevine@appen.com appen.com