Gen AI Cognizant & AWS event presentation_12 Oct.pdf
- 1. Beyond the hype: Harnessing the power of gen AI © 2023 Cognizant 12 October 2023 Philip Basford
- 2. 2 An AWS Partner since 2017 and Premier Partner since 2020 INAWISDOM As an all-in AWS business unit of Cognizant, our development and delivery teams live and breathe AWS services. Our team holds over 180 AWS certifications and accreditations. We maintain a close relationship with the AWS team, supporting and staying up-to-date with all the latest developments. Our team members hold individual certifications and accreditations in the following areas: ► ML Partner of the Year 2020 ► Global Launch Partner – Machine Learning ► Launch Partner – AWS UAE Region ► AWS Ambassadors We hold 9 competencies and service designations, reflecting business-wide expertise in key areas: Our Qualifications All our consultants hold at least 1 AWS certification. Including some consultants with all certifications Our CTO has been ranked #1 AWS Ambassador in EMEA for 2021 and 2022
- 3. The evolution of GENERATIVE AI Built on the last 30+ Years of progress Vast “vetted” corpuses are now available The Cloud has made huge amounts compute power available via on demand consumption Advances in AI architecture, especially on attention and transformers Simplification of use “We are at the iPhone moment for AI.” Jensen Huang, Chief Executive Officer, Nvidia
- 4. “Generative AI is neither a fad, nor an apocalypse, but Data & AI will power the innovation in business for the next decade.” Philip Basford, Chief Technology Officer, Inawisdom USE OF GENERATIVE AI Generative Search The ability to search a large amount of content and summarise the findings Smart Assistance for Data Analytics The ability to help the business to interact with their data and produce insights The possibilities with Generative AI are numerous, here are some examples: Advanced IDP The ability to summarise and extract content or data points from verbose inputs. Including grounded QA and RAG
- 5. USE OF GENERATIVE AI (ADDITIONAL) Developer Assistance Using Code Whisperer creates “boiler plate” code so developers can focus on business logic. Personalisation The ability to generate hyper-personalised experiences or marketing messages for an individual that represents a brand or product. Simulation The ability to create 3D models from images of infrastructure or buildings. In order to simulate large projects or the affect changes on the real-world including ESG impact. Routine Tasks The automation of routine tasks using Smart Assistants. This includes the assistant talking to or emailing other humans to order products or book events.
- 6. 6 Generative AI is a game changer, enabling increased efficiency and more innovation Over the next 5 years, Generative AI will become endemic within our lives Businesses are under pressure as competitors begin adopting Generative AI to gain competitive advantage Technology directors are struggling with the pace of change and new capabilities Enterprises cannot respond quickly enough compared to start-ups and risk-takers Face the potential leakage of data and Intellectual Property from unauthorised usage Business Challenges © 2023 Cognizant | Private Pace of Change and Disruption
- 7. 7 With Gen AI making headlines - both "good news" and "horror" stories - enterprises are wary of attracting any undue attention. As future regulation comes into play, solutions developed today may fall out of compliance Businesses fear being exposed to legal action or falling foul of copyright legislation Due to current economic pressures, budgets are restricted, and businesses struggle to know how to deliver maximum value from their AI investments. Many businesses struggle to understand what data they have, how to leverage it and how to get started with AI. Business Challenges © 2023 Cognizant | Private Controls and Responsibility
- 8. © 2023 Cognizant 8 Generative AI Architectures on AWS
- 9. Stable Diffusion • Generation of unique, realistic, high- quality images, art, logos, and designs Claude + v2 • LM for conversations, question answering, and workflow automation systems Jurassic-2z • Multilingual LLMs for text generation in Spanish, French, German, Portuguese, Italian, and Dutch Titan • Text summarization, generation, classification, open-ended Q&A, and search • Built 20 years of experience RAMP provides secure access to the widest range of FM in AWS FOUNDATIONAL MODELS ON AWS Command & Embed • Text generation model for business applications and embeddings model for search, clustering, or classification in 100+ languages Hugging Face • Repository of Open Source LLM and GPT models • Most models use Transferred Learning to refine models • Optimized Docker images and framework for distributed training Use Cases & Capabilities Sourced from AWS Amazon SageMaker • A full ecosystem for Machine Learning • API or Batch consumption • Pay per Min/Hour pricing • SageMaker has access to latest hardware including inf2 & Trn1 • Inawisdom has access to a wide range of FMs (proprietary + open source) • Inawisdom has worked with AWS at becoming specialists in distributed training. Initially using Hugging Face Amazon Bedrock • Managed Service for proprietary FMs • Proprietary FMs require EULA with FM Author • NEW: Agents for LangChain • FMs can be Fine-Tuned on your own data without you sharing your data with everyone • Currently in preview, access needs application • API based consumption (prompt+ completion style) + Pricing TBC New Service: Flagship Service:
- 12. Enterprise Knowledge Navigator “Please give me the current share prices for 10 best performing FinTech companies in the past 5 years and summarise their performance ” Advance Search / QA The ability to search inside private document, images or websites to find related content and then returning that content. Retrieval-Augmented Generation Integrations with live systems to augment the results with up-to-date information or perform actions may be required Security & Privacy Private FMs are not like Internet SaaS Products, your data is not shared and is kept securely
- 13. Enterprise Knowledge Navigator : Data Lakes The ability to help the business user to interact with their data lakes and produce insights Benefits: • Quick access of data to explore key insights or generate new insights from the data lake No SQL expertise needed in writing a good SQL • ~60-70% productivity gain, ask question in natural language and let generative AI (FMs) to do rest of work in generating insights for you Conversational Interface Providing a simple interface that allows the business users to speak/chat in plain English using domain specific phases. . Code and Domain Understanding Creating domain specific code to retrieve information contained within Data Products within a Data Mesh . Outcome Playback Generation of reports or a playback, containing generated graphics and text summarizing the result.
- 14. Enterprise Knowledge Navigator: Data Lakes
- 15. © 2023 Cognizant 15 Case Studies
- 16. Proven path to Gen AI The customer engaged to form a 5-year Gen AI roadmap. Cognizant (Inawisdom) is now validating the roadmap and looking at each use cases feasibility. The customer has selected Cognizant (Inawisdom) to prototype 2 use cases and pilot another A utilities customer ran a tendering process between a leading management consultancy and Cognizant, to select their Gen AI partner for a 4-year roadmap. Cognizant was selected as the preferred partner by leveraging Inawisdom's Gen AI knowledge and AWS strong relationship. In addition, Cognizant was able to add wider capabilities by bringing in relevant Subject Matter Experts (SMEs) from across the business, with expertise on sustainability and utilities. The customer are a leading private equity firm targeting technology buyouts primarily in Europe and the US. They are working with Cognizant (Inawisdom) to appraise their entire portfolio of companies and looking at how evolve their products with Gen AI to increasing the valuation of portfolio of companies © 2023 Cognizant | Private 16
- 17. Extraction of Data Used as part of IDP to extract structured information from text and images. Examples are invoice line items or complex nested data points where the relationship between them holds meaning Text Summarisation Generates new text that summarises the content contained from hundreds pages. This is typically used to pull out the key terms from very verbose documents The ability to help the business understand what is contained in their unstructured or semi-structured data IDP+ Text Classification The ability to look over the entirety of a piece of content or document to understand the type or use of the document
- 18. CASE STUDY IDP - From document-led to a data-driven marketplace The Customer: The Result: The Solution: The Requirement: Ø Trained & deployed fine-tuned LLMs targeted at domain specific documents Ø Established an automated, scalable underwriting process to improve underwriters’ day to day operations and drive business growth Ø Created intelligent AI solution to extract key data points (pricing/policies) from broker documents held in multiple types (pdf, email, xls) Ø Enabling faster velocity and quality for risk writing, encompassing various components and personas, to drive profitable business Ø Exploiting new innovations to improve accuracy in rating, forecasting, pricing and binding risk Ø Reducing operational costs Ø Creating a next-generation of market solutions to enable the business to be ‘future fit’ Ø Leading the digital revolution within the underwriting and risk process The Sector: Revolutionise the approach for underwriting risk in specialty insurance, leveraging AI & automated document processing Insurance International insurance and reinsurance group
- 19. 19 19 The Customer: The Result: The Solution: The Requirement: Ø Deployed custom ML models – using AWS SageMaker, Lambda and Step Functions – to interpret industry terminology and extract key data Ø Trained a classification model to detect potential errors in invoices and categorize them based on the primary reason for rejection Ø Leveraged Generative AI (GPT-3) to generate synthetic data for improved training and testing Ø Built a robust QA process and audit trail to ensure consistency and transparency Ø Accuracy rates of 75-97% across both use cases Ø 20% reduction in processing times Ø Yearly labour cost-savings of approximately $1.4m The Sector: Automate the summarisation of legal counsel guidelines and reduce errors during the invoicing process Business Services Provider of legal business and admin support services CASE STUDY Automating document processing & billing
- 20. 20 20 The Customer: The Result: The Solution: The Requirement: Ø Created a scalable document processing pipeline to extract key data from emails sent by brokers Ø Fine-tuned Large Language Models (LLMs) on AWS to extract and interpret industry-specific terminology Ø Developed a user interface to allow the underwriting team to review and correct the extracted data points as needed Ø Accuracy rates of 80-90% Ø Average processing time of less than 3 minutes, 540 times faster than the previous manual approach Ø Easy-to-use platform, with ongoing model improvement driven by underwriters’ feedback The Sector: Optimise the triage process for incoming leads to improve prioritization and speed up time-to-quote Insurance Specialty insurer underwriting personal & commercial risk CASE STUDY Accelerating lead processing in insurance
- 21. 21 21 The Customer: AI in Action: Optimising document processing in FSI The Result: The Solution: The Requirement: Ø Conduct remediation activities to improve existing IDP solution, implementing best practices for monitoring, scalability and integration Ø Develop new classification and data extraction models to handle a variety of structured and unstructured Retail Annuities documents, including free-form customer letters and application forms Ø Produce synthetic data using Generative AI to support training and testing of models, in place of sensitive customer data Ø Provide ongoing support and management of the solution Ø Faster data extraction and improved accuracy, leading to a reduction in processing costs Ø Improved error detection resulting in fewer documents being rejected The Sector: Improve and expand the existing IDP solution, to enable key use cases including accelerated processing of insurance documents Financial Services Leading provider of asset management & life insurance CASE STUDY
- 22. © 2023 Cognizant | Private 22 Industry Knowledge Governance and Reasonability Deep Technical Knowledge v The ability to contextualize Gen AI to an industry v Understanding and experience of challenges and common friction points v Awareness of industry direction over the next 5 to 10 years v AI Policy creation and advise v Knowledge of regulation and compliance v Knowledge of ESG and the human impacts of AI v Experts in Prompt Engineering, Fine Tuning and Foundational Model customization. v The ability to leverage Foundational Models from the AWS and Cognizant's Cognitive Pro TM for prototyping and LLM Ops Why Inawisdom for Generative AI? Product Centricity v The creation or evolution of user experiences v The ability to manage product lifecycle and launch products Business Readiness v The ability to create solutions that embed Gen AI in business process and evolve operating models v The ability to advise on readiness for Gen AI and how to evolve legacy technology
- 23. © 2023 Cognizant | Private 23 Build an AI Strategy How can Inawisdom help… Enablement Provide both business and technical enablement to teams to better understand Gen AI and the impacts it can have Ideation Bring the business and IT together with Industry SMEs from Cognizant to inspire big picture thinking and creation of a vision for AI and concepts for use cases Policy Construct an AI Policy on the usage of AI including what is prohibited and what is not. Scoring Down select and prioritize concepts by scoring them in terms of business value and complexity to deliver. Roadmap Take the scored concepts and design a roadmap that delivers the AI vision in accordance with the AI Policy. Unlocks incremental value with incremental investment at every turn
- 24. © 2023 Cognizant | Private 24 And execute it! How can Inawisdom help… Essential Controls From the AI policy implement the essential controls needed to initially start executing the roadmap Discover Validate concepts on the roadmap by creating the business case, likely return on investment (ROI) and success factors. Including looking at the feasibility of AI for the concept, running an EDA and checking the data readiness Prove Rapidly prototyping validated concepts and proving the value they can bring a business before further investment. Using the latest Foundational Models , AWS and Cognitive Pro TM Embed Creating pilot that is embedded within a business so that Success Factors can judged before full productionisation Adoption Transform a system, business process, or evolve an operating model to allow for Gen AI to reach its potential adoption and roll-out ….Then scale it!
- 25. AI & ML Flywheel © 2023 Cognizant | Private 25 Embed Maintain Evolve & Scale Data Sources Embed within business & visualise Structured, Semi-Structured and Unstructured data from Internal, External, and other sources Get stake holder commitment, build a roadmap around value and start the first flywheel for the highest impacting but deliverable use case Discover Business Case Creation, Exploratory Data Analysis, & Target Opportunity Definition Use Cases Prioritise & Value Business + Data Strategy, Ideation for Gen AI use cases Prove Experiment and show potential value Improve model(s), refine data products & create MVP Deliver value to the business Maintain value to the business Data & MLOps, maintain data & models with automation and pipelines 24/7 monitoring, Incident Response, & Cost Optimisation Respond to changes and detect drift Scale up with AIA and refine capabilities to accelerate the delivery of value with more & faster flywheels Improve reuse and collaboration using tooling such as a model registry and a Business Data Catalogue Standardise approaches to common problems, provide governance Change business processes and refine operating model to be data-driven POV with initial, data products, features creation & model selection Measure & Iterate Measure each iteration of the flywheel against CSFs / KPIs and only invest in further iterations as needed Roadmap Value Turning Hype into Reality
- 26. 26 Next Steps Get the latest insight in to Generative AI © 2023 Cognizant Learn how to practically apply Gen AI today for the greatest impact Hear from experts from across the technology landscape Get practical advice on getting started and unlock the myths surrounding Gen AI Read our latest Gen AI Report Coming to your inbox soon…