Rws Holdings Plc · LSE:RWS

Rws — Investor Day (October 2023)

126 slides from Rws's investor day presentation, Oct 10, 2023

Slide 1: AI / Technology Teach-in 10 October 2023
AI / Technology Teach-in 10 October 2023
Slide 2: Slide 2
Slide 4: Our aims for today
Our aims for today Unpack RWS’s AI / Technology story – past, present & future Showcase the Explain how AI Illustrate how capability & Outline some & technology we see AI expertise in AI of our future AI are critical to contributing to & technology / technology RWS & our growth and across the develo
Slide 5: Speakers and agenda
Speakers and agenda Ian El-Mokadem Alex Abey Maria Schnell 2.00 Welcome; RWS & Content Technology Internal Technology Deployment of AI Ian, Thomas, Mihai, Thomas Labarthe Panel Q&A Ian El-Mokadem Matt, Alex Technology / AI Summary Overview Technology 3.50 Break demonstrations Thomas, Vasagi, Panel Q
Slide 6: Who we are
Who we are A unique, world-leading provider of technology-enabled language, content and intellectual property services; well-diversified, with many leading market positions Clear structural demand £47bn market size Attractive drivers markets >8,000 corporate clients Proprietary language in 159 count
Slide 7: Strong portfolio, unique production platform
Strong portfolio, unique production platform Language Regulated Language & IP Services Services Industries Content Technology • Patent translation and OPERATING DIVISIONS • Localisation solutions • Life sciences • Linguistic AI - neural filing to multiple verticals MT • Financial services • Search,
Slide 8: Our purpose and values
Our purpose and values UNLOCKING GLOBAL UNDERSTANDING X X = We always keep our We play as one team – We shape the future – We choose to be promises – to clients, with colleagues, clients combining the best of positive – using every colleagues and and partners people and technology experience to grow
Slide 9: Our AI & technology story
Our AI & technology story Ian El-Mokadem
Slide 10: Why talk now about AI to investors?
Why talk now about AI to investors? Capital Markets Day (Mar 2022): • Stressed its importance • Welcomed its impact • Confirmed its central place in the Group’s future We recognize the hype… …however there is a real opportunity Today • Walk through the opportunities • Put the experts in front of you
Slide 11: What we have said previously
What we have said previously Macro environment has been challenging General & sector-specific factors have adversely impacted us AI, however, has not been a headwind – it’s a growing opportunity Our AI & technology propositions are making a positive difference 11 © 2023 RWS
Slide 12: Technology evolution and market growth
Technology evolution and market growth Successive technology waves have improved efficiency/productivity and lowered price/cost-to-serve Localisation market growth Statistical Machine 2020s Translation (SMT) 2010s Terminology databases / electronic 2000s dictionaries LLMs / Generative AI 1990s Neura
Slide 13: How to approach AI?
How to approach AI? Our core convictions relating to AI in our RWS is very well-placed to be a industry beneficiary Essential to adopt AI and play a lead role 1 in shaping its use While their roles will evolve, humans will Opportunity to be the valued partner 1 to our clients on their AI journey 2 c
Slide 14: RWS positive demand drivers
RWS positive demand drivers Increasing ESG Changing Explosion of Continued Growth in AI / / regulatory globalisation data / content innovation automation requirements market 90% of the world's >3k rules/regulations Capex investment AI to contribute Value of global trade data was generated added annu
Slide 15: Our strategy recognised the role of AI and technology
Our strategy, launched last year, recognised the role of AI and technology in driving growth and efficiency Long-term relationships Cultural & technical expertise • Neural Machine Translation and Linguistic AI – Language Unique technology Weaver and Trados Growth and AI • Content Management - Tridio
Slide 16: AI enabling a dramatic increase in content
AI is also enabling a dramatic increase in content of all types, creating opportunity for RWS Using a unique combination of technology and human expertise, we support our clients to create, collect, transform and analyse, launch and manage content This helps our clients grow, by ensuring they are un
Slide 17: 20+ years' experience gives deep capability and a competitive moat
20+ years’ experience gives us deep capability and a competitive moat Established Thought R&D Responsible Internal enterprise- leadership investment player deployment grade products Trados >£34m* >60% >40 Data Tridion patents Privacy FY22 MT first Language >100 Weaver peer- 600 Security MTQE reviewe
Slide 18: SDL acquisition enabled RWS from a technology perspective
SDL acquisition in 2020 enabled RWS from a technology perspective All services divisions leveraging LXD and associated LXD access technology, particularly Language Weaver LXD efficiency Increasing proportion of volume through LXD IP Services shift into Huge shift in thinking; only possible because w
Slide 19: AI – RWS's right to win
AI – RWS’s right to win Enterprise- Data Deep Enviable Attractive grade creation & expertise client set partner products validation capability We believe RWS is better placed than others in the sector to win 19 © 2023 RWS
Slide 20: Orientation – we will explore each of these areas
Orientation – we will explore each of these areas Backed by longstanding experience & expertise, RWS already able to support clients through their AI journey EXPLORING AI BUILDING AI USING AI Tech Services TrainAI U S I N G Weaver Language Language A I Weaver Trados Train your AI Understand Deliver
Slide 21: AI & Technology Overview
AI & Technology Overview Thomas Labarthe President – Language & Content Technology
Slide 22: We will explore each of these areas
We will explore each of these areas Backed by longstanding experience & expertise, RWS already able to support clients through their AI journey EXPLORING AI BUILDING AI USING AI Tech Services TrainAI U S I N G Weaver Language Language A I Weaver Trados Train your AI Understand Deliver translation Ch
Slide 23: Slide 23
Slide 24: Technology has unlocked more sophisticated content
Technology has unlocked more sophisticated content As content types are unlocked, specialist human expertise and technical capability become more critical Localisation market growth Content type commoditised – documentation 2020s Content type unlocked – web & social 2010s 2000s Content types unlocke
Slide 25: Terminology
Terminology A programme that can sense, reason, act and adapt Artificial Intelligence Algorithms whose performance improves as they are Machine exposed to more data over time Learning Subset of machine learning in which multilayered neural Deep networks learn from vast amounts of data Learning AI te
Slide 26: Example of LLMs being considered
Example of LLMs being considered Infrastructure Models Data Generic public • GPT 3.5 domain data • GPT 4 • Ability to generate data • GPT 3.5 Enterprise • Always • GPT 4 responsible, always ethical Enterprise • Cohere • Providing security & Private Open privacy Source LLM • Falcon 26 © 2023 RWS
Slide 27: LLMs are key to our product roadmaps
LLMs are key to our product roadmaps To productise them for Enterprise customers, we need to balance 4 variables: Linguistic features Adaptability Security and control Cost of performance Leveraging Language Weaver’s experience of productising other neural-based models 27 © 2023 RWS
Slide 30: Session overview
Session overview Product Right to win Client need overview and market and case and use attractive- study cases ness The future Progress and since Capital leveraging Markets Day LLMs 30 © 2023 RWS
Slide 31: Transforming enterprise content, securely, using Linguistic AI
Transforming enterprise content, securely, using Linguistic AI Three core use cases, across multiple end verticals USE Secure Multilingual Faster CASES translation content localization portal intelligence aid Technology only REVENUE Technology only MODELS (SaaS / Term / Perpetual) Technology and ser
Slide 32: Broad range of multi-sector clients
Broad range of multi-sector clients Government and defence Finance, legal and On-premises secure translation solution life sciences for defence intelligence applications Enable secure internal communications between global offices Allow governments to translate vast quantities of highly secure conte
Slide 33: Client need
Client need Fast, cost-effective Enterprise-grade, Adapted and fine- translation secure, scalable and tuned for the specific elastic solution • Diverse selection of • Various level of adaptation: language combination • On-premise or cloud dictionaries, fluent • Easy to use, supporting terminology, f
Slide 34: Centralising multilingual processes saves global IT services firm 60%
Centralising multilingual processes saves global IT services firm 60% RESULTS This Fortune 500 company offers IT services, consulting and business process outsourcing in multiple languages globally. 60% back-office “We were excited to explore the possibilities offered by Language Weaver. From proces
Slide 35: RWS right to win
RWS right to win Enterprise grade product Security, scalability, pre- & post-sales support Data Creation and Validation Ability to create and validate data supporting model training ‘Expert-in-the-loop’ Deep experience in AI / LLMs has allowed Language Weaver to move fast 35 © 2023 RWS
Slide 36: Model adaptation and upselling services are key to our growth
Model adaptation and upselling services are key to our growth Upsell More Services More Customers Revenue Feedback Loop Increased More Stickiness Data Higher Adaptation Win Rates Improved Generic Quality Invest in Partners and Extensions 36 © 2023 RWS
Slide 37: Progress since Capital Markets Day
Progress since Capital Markets Day Product • Neural Machine Translation (NMT) quality increased • Started rolling out MT Quality Estimation (MTQE) in LxD to increase post- editing speed • Rapid productisation of private LLMs is increasing the performance of NMT and MTQE – with 30% of the LW R&D focu
Slide 38: Continuous quality improvement
Continuous quality improvement t Generic Fluent Adaptable Quality sas | NMT 2.0 | Terminology | Models | Estimation Post Editing Automation RWS + Language Weaver
Slide 39: Linguistic AI developed significantly over past 7 years
Linguistic AI developed significantly over past 7 years Large Language Models (LLM) Bidirectional Representation Transformers (BERT) Transformers Single NLP task Multiple NLP tasks Multiple NLP tasks Single language Single language Multilingual 2016 2019 2023 39 © 2023 RWS
Slide 40: Different linguistic tasks require different AI models
Different linguistic tasks require different AI models EN FR DE RO FI PT ES IN MULTILINGUAL AR RU ZH … MULTILINGUAL Improve Input Adaptive MT Alternative Translations Smart MT EN FR MULTILINGUAL MT Quality Multilingual Estimation Summarization Beyond MT 40 © 2023 RWS
Slide 41: The future for Language Weaver
The future for Language Weaver Q1 FY24 • Enterprise-grade, private LLM for 2 use cases: • Improving MT generic quality through better input and output processing • Further improving localization, expanding the MT and MTQE technologies Beyond • Expand from document translations to document transforma
Slide 42: Key takeaways
Key takeaways Well Strong & Ongoing & LLMs give positioned in secure critical role opportunity a high enterprise- supporting to further growth grade RWS internal develop & market product efficiency improve 42 © 2023 RWS
Slide 43: Translation productivity and management
Translation productivity and management Matt Hardy Vice President, Product 43 © 2023 RWS
Slide 44: Session overview
Session overview Right to win Product Client need & market overview & & case study attractive- use cases ness Progress The future - since Capital Trados & AI Markets Day 44 © 2023 RWS
Slide 45: We serve three connected but distinct use cases
We serve three connected but distinct use cases Individual Collaboration Workflow & process USE CASES translation between teams management CUSTOMER SEGMENTS Corporate 1,700+    LSPs 350+   Linguist 250,000+  RWS internal    TRADOS Enterprise PRODUCTS Studio Team Accelerate 45 © 2023 RWS
Slide 46: We serve three connected but distinct use cases
We serve three connected but distinct use cases Individual Collaboration Workflow & process USE CASES translation between teams management CUSTOMER SEGMENTS Corporate 1,700+    LSPs 350+   Linguist 250,000+  RWS internal    Bundled TRADOS PRODUCTS Technology only with services Technology on
Slide 47: The Trados ecosystem
The Trados ecosystem Language Individual Service Commercial Translators Providers Enterprises Connect to Large Extend with Language Models Terminology Translation AI memory App Store Apps and Machine APIs translation providers Reporting Business systems Partner Workflow Translation solutions tools C
Slide 49: Client need
Client need Standardized & Value for all roles Easy integration of all efficient assets • Flexible ecosystem allows • Integrated workflow • Single solution deployment of any asset: solution for end-to-end encompasses all roles in • Machine translation management of translation the localization proce
Slide 50: LearnUpon localizes its user interface into nine new languages in three months
LearnUpon localizes its user interface into nine new languages in three months RESULTS “Look for a translation partner, not a translation company. We’ve built a strong bond with RWS and they’re playing an active role in helping us grow our business. That’s Quotes now provided what you want in a tran
Slide 51: RWS right to win
RWS right to win Established player with brand equity and clear suite of solutions Efficient & scalable development from multi-tenant cloud platform High levels of stickiness once clients have implemented technology choice 51 © 2023 RWS
Slide 52: Progress since Capital Markets Day
Progress since Capital Markets Day Embracing emerging AI technology Maintaining market leadership • OpenAI app for Trados Studio • Retain & enhance rich feature set (Freelancer BYO & customer LLM) • In-platform communication • LLM integration to Trados Enterprise • Increased workflow flexibility • A
Slide 53: The future - Trados and AI
The future - Trados and AI Q1 FY24 • Further AI features iterating Generative Translation Engines and Trados copilot • Delivering via responsible use of LLMs for customer data • Thought leadership through 15+ accepted conference presentation submissions • Subscription platform launch for Self Servic
Slide 55: Trados - key takeaways
Trados - key takeaways Trados retains Successfully pivotal role in Clear roadmap migrating localization in Maintaining to incorporate clients towards an AI-driven market LLMs & more latest world leadership AI-centred technology combining features platform HI + AI 55 © 2023 RWS
Slide 57: Session overview
Session overview Portfolio and In-depth Market AI and the category customer use perspective future overview case 57 © 2023 RWS
Slide 59: Content Technology
Content Technology Business-critical information Enterprise software produced by organisations 59 © 2023 RWS
Slide 60: Our category: component content management
Our category: component content management Author Manage Collaborate Publish Sales and marketing Product development Training and learning Claim management Accessible to all Omnichannel Modular content Stored centrally departments publishing 60 © 2023 RWS 60 © 2023 RWS
Slide 62: Documents to components
at A al L Documents A> omponents nt y LLL 4
Slide 63: Why content components?
Why content components? Write once Reuse many times Omnichannel publish Full audit trail 63 © 2023 RWS
Slide 64: What kind of content?
What kind of content? Hero marketing Content ideally content, reports, suited to LLMs as High value corporate component communications content co-pilots management Social content, Static LLMs as Low value email, outbound marketing content documentation substitute technology Transient Persistent 64 ©
Slide 67: Owner's Manual
Owner's Manual Choose your locale: North America Europe APAC Middle East English (United States/Canada) Catala (Catalan) ‘#02 (Mainland China) (AE) Francais (Canada) Espafiol (Spain) English (Mainland China) English (U.A.E.) Espajiol (Mexico) Hrvatski (Croatia) ig (Macau) aie (Jordan) Deutsch (Unite
Slide 68: Owner's Manual
Owner's Manual Manuale d’'uso Manuale d'uso Utilizzo del presente manuale d'uso < Panoramica Apertura e chiusura Vani portaoggetti Sedili e sistemi di ritenuta Connettivita Guida Autopilot Funzioni di sicurezza attiva Dashcam, sentinella e sicurezza Climatizzatore Navigazione e intrattenimento Ricar
Slide 69: Vehicle identification content
69 © 2023 RWS Vehicl le manufacturer Vehicle’s brand, body style, model, series, etc. Model year Vehicle’s serial number 4 S 3 BMHB6 8 3 286050 Country of manufacture Vehicle type Security check number Assembly plant 7 RWS
Slide 71: Customer support and aftermarket service
Customer support and aftermarket service
Slide 72: Customised repair manuals and instructions
Customised repair manuals and instructions
Slide 73: Employee training and dealership experience
Employee training and dealership experience
Slide 75: Web and mobile
7 = -\ a AVAILABLE IN 2 WEEKS Air Pure Lease at $749/mo. @ Buy at St “Destination and docun Rerraee 0-60 mp! 480 hp 38sec EPAest. range! Model Year A410 mi Dual Motor, All-Wheel Drive Stellar White Moj 19" Aero Range Wheels e PurLuxe with Alcantara Headliner Interior Platinum Appearance Metal Roof D
Slide 76: In-vehicle digital experience
In-vehicle digital experience
Slide 77: Legal and regulatory submissions
Consumer protection Autonomous vehicle Across jurisdictions Legal and regulatory submissions
Slide 78: Cars as software: updates = updated content
Cars as software: updates = updated content
Slide 79: And do all of this in multiple languages
And do all of this in multiple languages Tridion + Trados + Services 79 © 2023 RWS
Slide 80: Client need
Client need Process efficiency User experience Assurance Gives organisations a full Allows content to be Helps maintain solution to author, made accessible to the compliance manage & publish their right person, at the right content time, in the right medium Supports auditable chain & in the right c
Slide 81: Growth segments beyond manufacturing
Growth segments beyond manufacturing Life Sciences Financial Hospitality & Audit & Standards & Services Franchise Accounting Publishing 81 © 2023 RWS
Slide 82: RWS Content Technology portfolio
RWS Content Technology portfolio Rule makers Technical and Structured Aerospace & Use cases and rule regulated content Defence takers content authoring Products Propylon Tridion Fonto Contenta Focus Integrate & Continue cloud- Grow in Grow grow first journey & Pharma international grow markets 82 ©
Slide 83: RWS right to win
RWS right to win Established brands Major segments addressed, with minimal overlap Vertical-specific solutions and focus Tailored solutions for key markets like Pharma End-2-End solution Integration with localisation technology & services 83 © 2023 RWS
Slide 85: Key takeaways
Key takeaways Cloud-first Multi-product Targeting AI as co-pilot product portfolio vertical use drives roadmap will serves all cases beyond efficiency & allow client traditional productivity mid-market segments documentation growth 85 © 2023 RWS
Slide 87: Welcome back
Welcome back Ian El-Mokadem
Slide 88: Orientation – we will explore each of these areas
Orientation – we will explore each of these areas Backed by longstanding experience & expertise, RWS already able to support clients through their AI journey EXPLORING AI BUILDING AI USING AI Tech Services TrainAI U S I N G Weaver Language Language A I Weaver Trados Train your AI Understand Deliver
Slide 89: TrainAI / data services
TrainAI / data services Vasagi Kothandapani Senior Vice President, Strategic Accounts & TrainAI 89 © 2023 RWS
Slide 90: Every day many of us…
Every day many of us… Check our Use Maps Watch a Type an email Social Media to find movie using suggested feed our way recommendation text All of these are examples of AI How does AI learn? It needs vast amounts of data 90 © 2023 RWS
Slide 91: But what if the wrong data is used to train AI?
But what if the wrong data is used to train AI? This example might be funny but using incorrect data can have serious consequences AI says it’s AI says there’s a chihuahua. a threat. This is where TrainAI by RWS comes in – we provide clients with responsible AI training data 91 © 2023 RWS
Slide 92: TrainAI provides a broad range of data services…
TrainAI provides a broad range of data services… RLHF / Human-in-the-Loop QA Voice Domain expertise assistants Red teaming Natural language understanding Chatbots Content moderation Social media post classification Natural language generation Semantic segmentation Bounding box annotation Voice-contr
Slide 93: …including services to train and fine-tune generative AI
…including services to train and fine-tune generative AI DE ES Domain Data/content RLHF Risk Language expertise creation (reinforcement learning mitigation support from human feedback) • Subject-matter • Prompt engineering • Response rating, • Red teaming or • Locale-specific experts across a (promp
Slide 94: Here's the process
Here’s the process Raw AI data TrainAI Collect and/or Validate Deliver data specialists platform annotate data data training data SmartSource AI community of 100,000+ annotators, linguists, researchers, covering 400+ language variants across 175+ countries 94 © 2023 RWS
Slide 95: What does data collection look like?
What does data collection look like? Collection of text, audio, image, video data Prompt engineering Intent variations Images/videos Speaker, go Speaker, turn on Speaker begin Speaker play Turn on speaker Turn speaker on Start speaker Speaker, start Audio/speech recordings For illustrative purposes
Slide 96: What does data annotation look like?
What does data annotation look like? Text annotation Audio annotation Image annotation Video annotation PERSON DATE Ada Lovelace (1815-1852) was the first computer programmer, noted for her work on Charles Babbage‘s Analytical PERSON Engine. 96 © 2023 RWS
Slide 97: What does data validation look like?
What does data validation look like? Human-in-the-loop data QA Response QA, content moderation, fact verification, red teaming Individual review Consensus review For illustrative purposes only. Rater’s annotation manually Raters annotate same data point, checked by an expert and annotations are comp
Slide 98: TrainAI fulfils client needs for reliable AI training data
TrainAI fulfils client needs for reliable AI training data • Efficiency and productivity gains AI helps organisations deliver more • Improved speed of business • New capabilities efficient, tailored user experiences to • Business model expansion their customers and users • Increased profitability •
Slide 99: Big tech fine-tunes generative AI with TrainAI domain experts
Big tech fine-tunes generative AI with 200 TrainAI domain experts Client Objectives TrainAI solution Results • Boost LLM usability with Deliver the following based • Tight 4-week domain expertise on responsible AI principles: project ramp-up • Improve model safety • Recruiting and training • 200+ do
Slide 100: TrainAI has a compelling right to win
TrainAI has a compelling right to win Established capability 100,000+ strong community of AI data specialists covering 400+ language variants and 175+ countries Longstanding experience and reputation We have been supporting major technology firms for 10+ years Familiar route to an unsaturated market
Slide 101: Productising TrainAI
Productising TrainAI Early 2022 Apr-Sept, 2023 Ongoing Commitment to Data GenAI service offering developed Dedicated marketing initiatives Services Growth • Prompt engineering • Demand generation Opportunity • Reinforcement learning from human • Cross-sell/ upsell campaigns feedback (RLHF) • Sales s
Slide 102: Future developments
Future developments Go to market with GenAI Solidify TrainAI Enhance TrainAI launch and service extension operational model technology platform • Launch GenAI service offering in early 2024 • Enhance service delivery with AI • Continue to improve platform tools to drive data services industry expert
Slide 103: Key takeaways
Key takeaways Experience and Strong market Nimble and Revenue capabilities growth responsive opportunities Significant A high-growth Ability Opportunity to experience & proven industry with ready to respond quickly grow new revenue capabilities - access to clients to meet client streams beyond platf
Slide 104: Internal AI deployment
Internal AI deployment Maria Schnell Chief Language Officer
Slide 105: Session overview
Session overview LXD - RWS’s Localisation Efficiency production workflow opportunity platform Linguists Consequences differentiate for linguists us 105 © 2023 RWS
Slide 106: Language eXperience Delivery (LXD) - production platform
Language eXperience Delivery (LXD) - production platform Much more than Much more than Global market Expertise & translation… text-to-text… coverage capability • DTP • Image • 34 countries across • 29,000 freelancers • Engineering • Multimedia 5 continents with • 1,600 in-house • Testing • Voice – h
Slide 107: Performance and trends
Performance and trends 24hrs Av. turnaround time/project 63% Content going into LXD pre-translated by MT 1.9bn Words translated 68% of Language Services revenue Made up of: 1m Projects completed 60% of RI revenue 2022 LXD is already AI-dominant Language mix Increased demand for rarer languages with
Slide 108: Workflow – machine-first since early days of neural MT
Workflow – machine-first since early days of neural MT PDF = EN DE + + Client review Content Languages Quality Speed AI Client requestor Preparation Job Setup Translator Delivery to client Enabled by Human Adaptation Supported by Helix & 108 © 2023 RWS
Slide 109: What potential for further automation and productivity gains does AI bring to the LXD?
What potential for further automation and productivity gains does AI bring to the LXD? 109 © 2023 RWS
Slide 110: AI in Pre-Translation Stage
AI in Pre-Translation Stage Preparation Job Setup What we do today: What can still be optimised: Pre-translation from TM Improve input NMT Pre-translation Agile workflows with MTQE Resource recommendation Workflow recommendation 110 © 2023 RWS
Slide 111: AI in Translation Stage
AI in Translation Stage Translator What we do today: What can still be optimised: Autopropagate Research Autosuggest Quality Assurance 111 © 2023 RWS
Slide 112: AI in Adaptation Stage
AI in Adaptation Stage Adaptation What we do today: What can still be optimised: Automatic Speech Recognition Script translation Synthetic voice Voice mixing & cloning Automation first layout Automation first sub & dub Digital avatar Testing automation 112 © 2023 RWS
Slide 113: Complex adaptation – eLearning example
Complex adaptation – eLearning example PDF = EN DE + + Client review Content Languages Quality Speed Client requestor Preparation Job Setup Translator Delivery to client Adaptation Adaptation stage will include multiple rounds of reassembly and client review Layout Voice creation Reassembly Testing
Slide 114: Adaptation – what we have to bear in mind
Adaptation – what we have to bear in mind Human sentiment: Distrusting AI-generated content Ethical concerns: likeness, image & voice Need for balance between… …what technology will allow and… …what humans will accept from AI-generated content 114 © 2023 RWS
Slide 116: Two localisation pathways have evolved
Two localisation pathways have evolved Translation = Hyperpersonalization = Artificial Intelligence Human Intelligence • Source equals target • Relevant to target group, market, channel • Accuracy, consistency, compliance key • Native experiences are key • Goal: Risk reduction • Goal: Engagement • R
Slide 117: What the future might look like
What the future might look like Artificial Intelligence Human Intelligence • Pre-translate • Fact check & validate • Adapt language flavours • Transcreate • Supercharge: • Culturally adapt • Quality assurance • Make content accessible • Research • Optimising language assets • Optimise AI: • Longtail
Slide 119: Key takeaways
Key takeaways AI already Strong AI Production is plays critical opportunity at increasingly role to support all stages of automated efficiency production Growth and Translators will efficiency become benefits can be Language unlocked Specialists 119 © 2023 RWS
Slide 120: Closing comments
Closing comments Ian El-Mokadem
Slide 121: What we shared today
What we shared today Unpacked RWS’s AI / Technology story – past, present & future Showcased the Explained how Illustrated how capability & Outlined some AI & we see AI expertise in AI of our future AI technology are contributing to & technology / technology critical to RWS growth and across the dev
Slide 122: Our core convictions relating to AI in our industry
Our core convictions relating to AI in our industry Essential to adopt AI and play a lead role in shaping its use While their roles will evolve, humans will continue to play a key role in content transformation AI will continue to enable further improvements in cost per word and this will be balance
Slide 123: RWS's right to win
RWS’s right to win Enterprise- Data Deep Enviable Attractive grade creation & expertise client set Partner products validation capability We believe RWS is better placed than others in the sector to win 123 © 2023 RWS
Slide 124: The key ingredients
The key ingredients FOCUS ON CONTENT LEVERAGE EMBRACE TRANSFORMATION GLOBAL SCALE ARTIFICAL INTELLIGENCE AND REACH AND THROUGH LXD HUMAN INTELLIGENCE 124 © 2023 RWS
Slide 125: Genuine Intelligence
GENUINE INTELLIGENCE When AI and HI combine, you create a world of GI. Genuine Intelligence. 125 © 2023 RWS

RWS Holdings — Investor Day (October 2023) on 6ix

Rws — Investor Day (October 2023)

Slide 1: AI / Technology Teach-in 10 October 2023

Slide 2: Slide 2

Slide 3: Welcome, introduction, objectives

Slide 4: Our aims for today

Slide 5: Speakers and agenda

Slide 6: Who we are

Slide 7: Strong portfolio, unique production platform

Slide 8: Our purpose and values

Slide 9: Our AI & technology story

Slide 10: Why talk now about AI to investors?

Slide 11: What we have said previously

Slide 12: Technology evolution and market growth

Slide 13: How to approach AI?

Slide 14: RWS positive demand drivers

Slide 15: Our strategy recognised the role of AI and technology

Slide 16: AI enabling a dramatic increase in content

Slide 17: 20+ years' experience gives deep capability and a competitive moat

Slide 18: SDL acquisition enabled RWS from a technology perspective

Slide 19: AI – RWS's right to win

Slide 20: Orientation – we will explore each of these areas

Slide 21: AI & Technology Overview

Slide 22: We will explore each of these areas

Slide 23: Slide 23

Slide 24: Technology has unlocked more sophisticated content

Slide 25: Terminology

Slide 26: Example of LLMs being considered

Slide 27: LLMs are key to our product roadmaps

Slide 28: © 2023 RWS

Slide 29: Language Weaver

Slide 30: Session overview

Slide 31: Transforming enterprise content, securely, using Linguistic AI

Slide 32: Broad range of multi-sector clients

Slide 33: Client need

Slide 34: Centralising multilingual processes saves global IT services firm 60%

Slide 35: RWS right to win

Slide 36: Model adaptation and upselling services are key to our growth

Slide 37: Progress since Capital Markets Day

Slide 38: Continuous quality improvement

Slide 39: Linguistic AI developed significantly over past 7 years

Slide 40: Different linguistic tasks require different AI models

Slide 41: The future for Language Weaver

Slide 42: Key takeaways

Slide 43: Translation productivity and management

Slide 44: Session overview

Slide 45: We serve three connected but distinct use cases

Slide 46: We serve three connected but distinct use cases

Slide 47: The Trados ecosystem

Slide 48: LIVE DEMO

Slide 49: Client need

Slide 50: LearnUpon localizes its user interface into nine new languages in three months

Slide 51: RWS right to win

Slide 52: Progress since Capital Markets Day

Slide 53: The future - Trados and AI

Slide 54: © 2023 RWS

Slide 55: Trados - key takeaways

Slide 56: RWS Content Technology

Slide 57: Session overview

Slide 58: Content Technology portfolio

Slide 59: Content Technology

Slide 60: Our category: component content management

Slide 61: What are content components?

Slide 62: Documents to components

Slide 63: Why content components?

Slide 64: What kind of content?

Slide 65: Video

Slide 66: Use Case: Large EV manufacturer

Slide 67: Owner's Manual

Slide 68: Owner's Manual

Slide 69: Vehicle identification content

Slide 70: Reuse, Consistency

Slide 71: Customer support and aftermarket service

Slide 72: Customised repair manuals and instructions

Slide 73: Employee training and dealership experience

Slide 74: Brochures and collateral

Slide 75: Web and mobile

Slide 76: In-vehicle digital experience

Slide 77: Legal and regulatory submissions

Slide 78: Cars as software: updates = updated content

Slide 79: And do all of this in multiple languages