1. Public Being Streetwise with MT in an Enterprise Neighborhood Will Burgett, PMP® Product Manager Translation Innovation & Services Marketing Solutions November 4, 2010

2. Public Legal Notices 2 Translation Innovation & Services November 4, 2010 THIS PRESENTATION CONTAINS THE GENERAL INSIGHTS AND OPINIONS OF INTEL CORPORATION (INTEL). THE INFORMATION IN THIS PRESENTATION IS PROVIDED FOR INFORMATION ONLY AND IS NOT TO BE RELIED UPON FOR ANY OTHER PURPOSE THAN EDUCATIONAL. USE AT YOUR OWN RISK! INTEL MAKES NO REPRESENTATIONS OR WARRANTIES REGARDING THE ACCURACY OR COMPLETENESS OF THE INFORMATION IN THIS PRESENTATION. INTEL ACCEPTS NO DUTY TO UPDATE THIS PRESENTATION BASED ON MORE CURRENT INFORMATION. INTEL IS NOT LIABLE FOR ANY DAMAGES, DIRECT OR INDIRECT, CONSEQUENTIAL OR OTHERWISE, THAT MAY ARISE, DIRECTLY OR INDIRECTLY, FROM THE USE OR MISUSE OF THE INFORMATION IN THIS PRESENTATION. Intel, Intel logo, Intel Inside, Intel Inside logo are trademarks or registered trademarks of Intel Corporation or its subsidiaries in the United States and other countries. “PMP” and the PMP logo are certification marks of the Project Management Institute which are registered in the United States and other nations. *Other names and brands may be claimed as the property of others. Copyright © 2010, Intel Corporation. All rights reserved.

3. Public What is the Enterprise Neighborhood like Today? It’s global and very competitive –Even things that were once strictly local are not anymore –Healthcare It’s a tough neighborhood –As our former and famous CEO Andy Grove wrote and said, “only the paranoid survive.” It’s constantly challenged by change –Smart phones, Netbooks, PC tablets, the cloud It’s political It’s resource constrained –Natural- Budget –Human- Time 3 Translation Innovation & Services November 4, 2010

4. Public Streetwise with MT Streetwise means “having the shrewd resourcefulness needed to survive in an urban environment.” Quickly adapting to change as the neighborhood changes. A truly global economy means that disruptive forces are constantly at play, making the enterprise neighborhood dynamic and transitory. So with MT, not only do we have to reduce the cost of translated content, increase productivity, speed the time-to- market, but we also have to reach our global customers, build lasting relationships, maximize revenue while meeting a moving target for quality. 4 Translation Innovation & Services November 4, 2010

5. Public “Urban Planning” Comes First in the Enterprise Neighborhood It must be an Integrated Plan 5 Translation Innovation & Services November 4, 2010 Global Content Management Content Quality Management Workflow & Translation Automation Data Sharing

6. Public Pyramid of Content Dependence 6 Translation Innovation & Services November 4, 2010

7. Public The Enterprise Neighborhood The Traditional Triple Constraints and Improvements 7 7 Translation Innovation & Services November 4, 2010

8. Public Original Motivation at Intel to Move to Machine Translation in Customer Support Cost –Reached maximum cost efficiencies of human translation –Reached maximum budget levels for human translation of customer support content Time (Time-to-Market) –Took 10 business days for complete translation cycle even with workflow automation (1200-1500 files per project cycle) –Source and target always 10 business days out of sync Quality (Fully Automatic Useful Translation = FAUT) –MT quality has gotten better –Users more willing to accept language imperfections in exchange for useful information 8 Translation Innovation & Services November 4, 2010

9. Public Machine Translation Integrated into Intel Customer Support 9 Translation Innovation & Services November 4, 2010

10. Public Intel Customer Support – LAR Spanish 10 Translation Innovation & Services November 4, 2010

11. Public The Enterprise Neighborhood with Machine Translation 11 Translation Innovation & Services November 4, 2010

12. Public Results: Cost Savings, Productivity, Time-To-Market 2008-2010: Saved Intel millions of dollars in translation costs Decreased cycle time from two weeks to 24 hours [Intel Customer Support (ICS) with machine translation (MT)] Increased translation rate from 300 to 60,000 words/hour (human vs. MT) Reduced translation cost per average page by 95% (unedited MT for ICS) Enabled the Customer Support self-help model in other languages 12 Translation Innovation & Services November 4, 2010 Intel Customer Support budget decreased almost half between 2006-2010, yet translation volume grew 3.5X

13. Public Customer Acceptance of Fully Automatic Useful Information Intel Customer Support survey widget on every page “Did this information help answer your question?” Widget also collects direct customer feedback Results for the machine-translated sites: –About.01% of the comments relate to the translation Detailed survey also run at the same time which on a quarterly basis shows similar results 13 Translation Innovation & Services November 4, 2010

14. Public New Influences in the Enterprise Neighborhood 14 Translation Innovation & Services November 4, 2010

15. Public New Neighbors on the Enterprise Block Reach, Relationship, Revenue Reach –Deliver customer-oriented interactive Web experiences –Customer experience –Use the Web to reach customers in a dynamic interactive way –Remove the barriers of language and location  Depends on existing translation platform plus new multilingual interactive applications Relationship –Flexible collaboration makes Intel “Easy to do business with” –Multilingual internet, chat, Email, instant messaging, communities and forums, blogs –Intel’s major direct customers are Original Equipment Manufacturers, Original Design Manufacturers, Distributors, etc. –Customer and brand loyalty are major objectives  Quality of source and target languages important Revenue –80% of Intel’s 2009 revenue came from outside the North American Region –Faster time-to-market means faster time-to-revenue –Machine translation can deliver content closer to the time of purchase and support 15 Translation Innovation & Services November 4, 2010

16. Public Intel Revenue in 2009** 16 Translation Innovation & Services November 4, 2010 ** Source: Intel Corporation 2009 Annual Report

17. Public Content Disruption in the Neighborhood Content is growing dramatically: Average enterprise sees information increasing 50% to 100% per year. Publisher-centric model is fading as users take charge. –Instead of pushing out translations of largely unused documents, many organizations switching to “pulling in” users with specific information needs in the form of dynamic content. There is a huge tidal wave of user-generated content that is driving a significant shift in strategy. 17 Translation Innovation & Services November 4, 2010

18. Public A Sea Change through Social Media Social Media gives the audience a voice and control. Conversely, traditional media pushes a message, and the audience mostly observes and consumes. Social media can build relevance in a message because it involves the audience and a voice of many over the voice of a few. 18 Translation Innovation & Services November 4, 2010

19. Public Social Media is Connecting the Enterprise Neighborhood Center for Marketing Research, University of Massachusetts Dartmouth annual study on social media adoption and practice by the Inc. 500. 91% of companies reported use of at least one social media service or tool in 2009. Social Media is introducing a competitive advantage, with adoption ensuring survival and success through practice and evolution. 19 Translation Innovation & Services November 4, 2010 Source: The Fortune 500 and Social Media: A Longitudinal Study of Blogging, Twitter and Facebook Usage by America’s Largest Companies Nora Ganim Barnes, Ph.D. Center for Marketing Research at the University of Massachusetts Dartmouth

20. Public Social-media Platform Enables Intel’s Social Media Strategy Intel Social Strategy Objective Increase the relevance and preference of the Intel brand in the online social dialogue Enable & Amplify Voice of Influencers Move the voice of influence from Intel’s institutional voice by fostering, extending, enabling, and amplifying the voice of the following influencers online: –Intel’s own subject matter experts –Press and influential media –Ecosystem partners and fellow travelers –Community of advocates and customers But an old problem: Just as in software, documentation, the Web, you have to duplicate cost and effort to create language versions! 20 Translation Innovation & Services November 4, 2010 Intel’s Experts Media Ecosystem Community Advocates

21. Public 21 Translation Innovation & Services November 4, 2010

22. Public Example of Social Media and Machine Translation XIHA, Helsinki, Finland, Founded in 2007 XIHA developed world’s first multilingual social media platform (powered by Google Translate) and operates XIHA Life, an online community targeted at people who use multiple languages in their everyday lives. Has 700,000 members who come from over 200 countries Works with Facebook* and Twitter 22 Translation Innovation & Services November 4, 2010 *Other names and brands may be claimed as the property of others

23. Public Challenges for Research and Commercial Developers of Machine Translation to Keep the Enterprise Streetwise Social media and user-generated content –Empower the blogger to initiate machine translation and the community to respond –Post-editing tools for the community and crowd sourcing Unstructured, unconstrained, real time Multilingual interaction and collaboration Not just customer support, but also sales and marketing interactions Machine translation must not only reduce cost, increase productivity, and reduce the time-to-market, it must also help enterprises reach their customers to build relationships that maximize revenue and the customer experience. Oh, and quality is still very important 23 Translation Innovation & Services November 4, 2010