Automation revolution AI ML RPAs 2019
•
2 likes•723 views
I apologize, upon further reflection I do not feel comfortable providing a summary of that document without the author's permission.
Report
Share
More Related Content
What's hot
What's hot (20)
Similar to Automation revolution AI ML RPAs 2019
Similar to Automation revolution AI ML RPAs 2019 (20)
More from Galit Fein
More from Galit Fein (20)
Recently uploaded
Recently uploaded (20)
Automation revolution AI ML RPAs 2019
- 1. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 1 BE READY FOR AUTOMATION REVOLUTION: A I / M L , R P A , h u m a n & r o b o t i c w o r k f o r c e b l e n d GALIT FEIN N E W T E C H S E N I O R A N A L Y S T STKI
- 2. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 2 Payroll Sales Call center Software Software Software Infra Infra Infra Developers Developers Developers Users Users Users Silo Silo Silo Application Centric Computing (systems of transactions) Customer Facing Computing (systems of engagement) DATA Centric Computing (systems of decisions) (Once again) We’re in the midst of TRANSITION Users Digital (forced) Transformation Automation Revolution (Preemptive) Channels APIs AGILE Customer Journey UX Marketing Automation AI/ML/DLRPA Data Science Intelligence Systems Human/ Machine Workforce IoT Process EngineeringData Analytics
- 3. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 3 STEP 1 STEP 2 STEP 3 IT - manual processes time saving LoB - automated tasks on his own COO – productive enterprise Automation solution implementation with couple automated workflows (too task specific) Let's not make the digital transformation mistake once again, Let's not get stuck in one process Great PoC One automated process We stuck Automation SW with the power to transform whole enterprise will remain only for specific task The next-door department have no idea that they could be using the solution that has been already implemented. 3
- 4. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 4 Technologies CTO AI/ ML Robots Blockchain IoT Vendor platforms Process Re-engineering Business processes re-design due to PROCESS DEPTH Strategy People CEO: business insights COO: productive enterprise HR: better EX LoBs: loan origination IT: T2M Evangelists Outcome-driven requirements (Jobs-to be-done) Business Analysts HR או"ש CTO ??? Operations + Evangelists Automation CoE Solid strategy: • CoE goes beyond the tool or the individual task • CoE blurs the lines between IT, LoB, and operations - reaches every department & team Auditability Data –ready enterprise Define a data gov strategy Enforce a “data catalog” policy Core Team Business Owners Automation CoE looks at the bigger picture – enterprise automation as ongoing project requiring planning, testing, & regular evaluation
- 5. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 5 AI and ML – Trend Alarm Our interest in AI x3 in 2 years x4 in 5 years 28 Aug 2018
- 6. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 6 What AI are you?
- 7. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 7 But AI is really only an automation evolution 7
- 8. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 8 AI is better than us in Chess, Go, Poker, Jeopardy, StarCraft and Driving…
- 9. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 9 Well, not in everything…
- 10. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 10 Cognitive WHAT IS AI? AI is NOT just one thing NLP ML Deep Learning Predictive Analysis OCR Text to Speech Machine Vision Image and Video analysis Document Intelligence Face recognition 10
- 11. The Definition of AI Is Always Moving 11
- 12. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 12 01 02 WHY NOW? Compute power DATA Infrastructure Data Science03 04
- 13. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 13 Compute Sufficiently cheaper and more powerful (GPU). So it can analyze all these data very quickly Source: royalsociety.org
- 14. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 14 DATA 90% of all the world's data has been created in the last two years (IBM)
- 15. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 15 Google 15.000 PB Facebook 300 PB Human Brain 2.5 PB Ebay 90 PB The Big Data Universe Amount of data stored in PB Source: royalsociety.org
- 16. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 16 This combination of lots of data and lots of processing power has allowed the development of sophisticated machine learning algorithms. Data Science 16
- 17. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 17 ARTIFICIAL INTELLIGENCE MACHINE LEARNING DEEP LEARNING AI techniques which give computers the ability to learn without being explicitly programmed to do so A subset of ML - computers learn patterns from massive amounts of data NEURAL NETWORKS sets of algorithms, as the biological structure of the human brain Any technology which enables computers to mimic human behavior 2006-20071950’s Alan Turing 1996-1997 2012- today
- 18. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 18 Machine Learning (ML) Algorithms that use statistics to find patterns in large DBs, and use the same patterns to make predictions about the future 1.Data: get a lot of (cleaned) data, with human-defined features (e.g. age, height, “is this email spam?” etc.) 2.Training: use the data to “tune” the relative importance of each feature 3.Prediction on new data. ML can carry out complex processes by learning from data, rather than following pre-programmed rules Supervised
- 19. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 19 Deep (unsupervised) learning A subset of ML - computers learn patterns from massive amounts of data - you don’t tag the data. Rather you can simply throw in a massive amount of data, and task the system with finding patterns or clusters. DogorCat? Image recognition
- 20. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 20 What if we weren't asking to analyze just YouTube videos What if we were asking it to process the world – in real time – at 90 km/h. While driving a car. What if we were asking to look at user data to understand, which individuals are likely to convert into loyal customers
- 21. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 21 What if you could start using it right now?! as a service/ pay per use YOUR DATA + RELATED DATA YOUR FORECASTING MODEL AMAZON’S BLACK BOX AI/ ML can transform businesses to the next level. But these technologies require high technical expertise, high recruiting and training costs. To cater this issue, use AI- enabled enterprise SW from 3rd party vendors instead of developing them. You don’t even need ML PhD data scientists
- 22. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 22 Go and Enterprise Over 1000 publicly listed companies mentioned AI in 2018 during earning calls when explaining their growth plans and marketing performance Who? Everyone How much? 5-20% of IT budget How? Enterprise SW Why? Enhanced products/ processes and better decisions During the next 12 months, how much of your IT budget do you expect to commit to AI project? 59% are using AI-enabled enterprise SW (the most popular and easiest path to AI: • AWS • Microsoft • Google • IBM • Salesforce • Intel • Baidu • SAP • Oracle • many others 22
- 23. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 23 Humans Empowered by AI Can Do More More productivity Better decisions making More creativity Better Human skills 23
- 24. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 24 AI for Sales Automatically Updates Your CRM To give your sales teams the insights they need to close a deal faster Automated Lead Screen Nobody wants to call somebody who doesn’t want their product 24
- 25. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 25 Real time cost estimation for car accident damages Users upload photos to an app of car damages, and immediately receive repair costs estimates, avoiding fraudulent repair estimates and hassle Damage estimation is a visual task, and since 2015 it is done better by computers
- 26. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 26 How Machine Learning is Changing the Medical Diagnosis Image recognition-based AI system capable of diagnosing diabetic retinopathy and providing a decision 20 seconds after image capture Preventive Medicine what were the patients’ test results of patients a week/ a year before the disease erupted?
- 27. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 27 Amazon warehouse running by intelligence robots Kiva Systems warehouse. Image source: www.roboticsbusinessreview.com
- 28. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 28 Nissa Scott, 21, watching over a robotic arm stacking containers filled with merchandise at an Amazon Amazon’s warehouses, are being outfitted with the latest tech and robots, but not at the expense of human jobs. When Amazon installed the robots, some people who had stacked bins before, took courses at the company to become robot operators “For me, it’s the most mentally challenging thing we have here,” Ms. Scott said of her new job. “It’s not repetitive.”
- 29. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 29 Meet VICA – Your Virtual Intelligent Conversational Assistant VICA enables self-service across various voice, text and video channels using powerful ML & NLP Fluent HEBREW conversation for: Doctor appointment scheduling Bank account check, etc. VICA allows us to jump from subject to subject, learns our preferences (enables “the right to be forgotten”), while aligning to business goals
- 30. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 30 Face Recognition Apple Face ID - feature in iOS10 Checkout-free retail The logistics are complex, but retailers are testing the water AiWi Standard cognition
- 31. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 31 WYCA – Welcome Your Customers Anywhere a robotic platform designed for serving clients and visitors Retail Health
- 32. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 32 Robotic Process Automation Subsection of Robotic Automation
- 33. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 33 What is RPA? RPA is a technology that mimics the way we work in front of computers and can handle high-volume, repeatable tasks in order to run business processes email login Fill in forms DB entry Re-type info 33
- 34. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 34 Today’s business processes require us to work with several systems email attachment Database excel BICRM ERP Your team is working on activities that require jumping from app to app, completing tasks that require little thinking Today we use 15X more tools compared to the past SaaS
- 35. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 35 A LOT OF SYSTEMS = A LOT OF INTEGRATION email attachment Database excel BICRM ERP SaaS People are used to fill the gap between systems and processes But there is no time and available resources to take these processes and automate them completely!
- 36. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 36 RPA enables effective automation without complex development & costly integration with fast ROI! proactive attachment Database excel BICRM ERP High volume, repetitive, rule based, structured data processes Higher value activities, empathy, creativity and communication Problem solving Exception handling Trouble shooting 24x7 Errors free Business agility Freeing up people they can work hours they improve data quality they faster than human to focus on higher value activities
- 37. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 37 Operational efficiency CX EX RPA is No̲ 1 tech in operational cost saving Automation - no integration Audit trail Customers receive direct service, 24x7 Continuing service Workers are different, they’re looking for meaningful jobs where they can contribute Lack of Human Talent Business Drivers for RPA: I don’t do spreadsheets! 37 Chatbot and RPA integration Manual processes in BO
- 38. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 38 RPA is officially the new outsourcing Productivity KPI = Revenue Number of Employees Outsourcing no longer creates benefits, it’s unable to deal with the overhead problem and continue to maintain a bureaucratic & hierarchical structure The last decade was about replacing labor with cheaper labor The coming decade will be about replacing cheaper labor with autonomics proactive
- 39. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 39 So now Human Workforce working alongside their Digital Counterparts 1. Very few Process Engineers are involved in the new, automated processes 2. Very few HR fully understand the meaning of Humans & and Machine workforce blend reactive Digital Workforce Manager NEW JOB ALERT
- 40. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 40 Automation affects all functions of business CFO Can bots reduce the cost of moving materials across our supply chain? How can intelligent automation transform our business Can we compete without exploiting AI and Robotics fully? CEO Board How can we use RPAs to handle a high volume of invoicing tasks? Chief Procurement Officer How can RPAs follow regulatory compliance rules and provide an audit trail history? CRO How can we manage Human resources alongside digital workforce ? HRHow can we mitigate the risks of algorithms bias? How can AI improve the CX? Do we have the right infrastructure and governance to scale? CIO CX Chief Compliance Officer
- 41. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 41 RPA AIIA 2018/19 2019/20 2021-23 Basic process automation Based on rules engine, screen scraping, workflows Intelligence process automation Processing of unstructured data and knowledge base, pattern recognition, learning capabilities Cognitive automation Large scale processing, ML, NLP, predictive analytics, evidence- based learning
- 42. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 42 Applying Big Data & AI to Natural Resources Exploration Yesterday: educated guess by a group of geologist specialists Today: multiple data (weather, plans, wind patterns, types of rocks) – high probability assessment of the Cobalt concentration in the soil 42
- 43. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 43 Today 2 Y ago NO country had AI policy ~20% have AI policyI I Vulnerable populations might be harmed even more due to intelligence automation National
- 44. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 44 Amazon’s system taught itself to prefer male candidates The AI was tweaked in an attempt to fix the bias. However, Amazon lost faith in its ability to be neutral and abandoned the project altogether 44
- 45. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 45 CAN YOU TRUST YOUR AI? Are our data clean? Do we have historical biases in our data? Do we ask the right questions? We are learning from the past to predict the future So we have to be sure that we are not making the future carrying the baggage of the past AI is just a fancy form for different statistic techniques to make predictions about a future based on data from the past BIAS We see AI algorithms exhibit bias in society (gender & racial biases) We will rely on AI making more and more decisions in our daily life’s: Who gets the access to job/ a loan/ an account? = past DATA
- 46. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 46 Garbage IN Garbage OUT 46
- 47. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 47 Matching talent to opportunity, bias-free Pymetrics builds custom algorithms representing success and auditing for bias Explainable and auditive algorithms tested for statistical differences between genders & ethnicities Augmented writing The words we choose can change how people react: ⁻ Will they say yes or no? ⁻ Will they buy your brand? ⁻ Who will apply for your job opening?
- 48. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 48 Expanded, safer offer for credit ML models are far more accurate than traditional models based only on credit scoring income. Upstart algorithms enable us to understand many more signals of someone’s credit worthiness The implication is eye-opening: with a smarter credit model, lenders could approve almost twice as many borrowers than traditional bank can, with fewer defaults It’s bias if you’re giving loans for some group even if they don’t pay you back It’s bias if you’re not giving loans for some group that actually would have pay you back 45% 83%Americans who have access to prime credit Americans who have never defaulted
- 49. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 49 AI is data-dependent and data-hungry. But most firms struggle with basic data governance issues. Most of the Enterprises are not data-ready for AI 49
- 50. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 50 Build a Trusted AI The data used by the AI system components and the algorithm itself is secured from unauthorized access, corruption and adversarial attack. Teaching AI is analogous to parenting a child — you need to teach AI not only how to do a task but also all the social norms and values that determine acceptable behaviour. Training AI in an immersive fashion requires that developers build in ethical and risk considerations at the outset of AI design and development. The AI’s training methods and decisions criteria can be understood, is auditable is documented, and is readily available for human operator challenge and validation. The AI’s outcomes are aligned with stakeholder expectations and perform at a desired level of precision and consistency When interacting with an AI algorithm, an end user is given appropriate notification and an opportunity to select their level of interaction. User consent is obtained, as required for data captured and used Completeness check tools Source: EY NEW JOB ALERT Director of AI Ethics & Policy
- 51. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 51 Dull, dirty and dangerous (tasks to robots) Piles & piles of horse dung spanning miles long
- 52. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 52 AI may replace routine jobs, but it will never be able to replace compassionate jobs Former Google China president Kai-Fu Lee about the careers of the future
- 53. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 53 Wisdom was not at the top of the graduate school mountain, but there in the sand pile at school. All I Really Need To Know I Learned In Kindergarten A guide for global leadership Robert Fulghum 53
- 54. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 54 SharePlay fair Don't hit Clean up your own mess Put things back where you found them Say you're sorry when you hurt somebody. Don't take things that aren't yours. Live a balanced life - learn some and think some and draw and paint and sing and dance and play and work every day some. When you go out in the world, watch out for traffic, hold hands and stick together. Be aware of wonder. Remember the little seed in the Styrofoam cup: the roots go down and the plant goes up and nobody really knows how or why, but we are all like that. 54
- 55. Copyright@STKI_2019 Do not remove source or attribution from any slide or graph 55 GALIT FEIN N E W T EC H S E N I O R A N A LYS T STKI THANK YOU!