Provides IT services, consulting, and business solutions.pdf
- 2. AI for AI solutions engineering Sankaranarayanan Viswanathan and Suranjan Chatterjee 2
- 3. Digital transformation : Advent of Generative AI, scales the degree of automation Infuse RPA, traditional AI &ML to drive efficiencies and consistencies Digitize and standardize what is manual and slow Assist humans through automation… 3
- 4. Digital transformation : Advent of Generative AI, scales the degree of automation Infuse RPA, traditional AI &ML to drive efficiencies and consistencies Digitize and standardize what is manual and slow Engineered tech leveraging gen AI to improve the quality and speed of decisioning Paradigm shift from automation to influencing quality of outcomes 4
- 5. Core enterprise IT and OT Computing, network, data storage, digital products and platforms, and systems of record Third-party and partner-provided; open-source Existing enterprise IT and OT Multitiered architecture of future AI-first businesses Custom; AI engineering and systems integration Customers Employees Reimagine Ways of Working (WoW) and “jobs to be done” to drive new value Key to market differentiation Bringing it all together: TCS POV | AI-first business architectureTM Approach to transforming knowledge work Foundational LLM, DWH, and data lakes Structured data stores, unstructured data, external data stores, analytics, and insights Purposive and contextual task agents Create and tune models; set up guardrails; and observe, learn, and adapt AI-augmented work systems Intelligent orchestration of purposive task agents; human-in-the-loop 5
- 6. The new paradigm needs investments in three areas to unlock outcomes 6 New AI architecture that allows enterprises to harness AI across clouds, models, and providers Foundational enterprise AI fabric New fabric for knowledge worker transformation resulting in transformed enterprise outcomes Knowledge worker augmentation Transforming the knowledge value chain to deliver the right enterprise outcomes Value chain transformation Assist Augment Transform
- 7. Thank you Copyright © 2024 Tata Consultancy Services Limited 7
- 8. AI for Software engineering Muthulakshmi Nellaiappan 8
- 9. Our focus lies in building Building software is not just writing code that most AI tools claim today … Safe-to-use Easy to maintain software Enterprise ready Fit-for-purpose 9
- 10. Building enterprise-ready software 10 AI brings a world of opportunities; we have the responsibility to do it right Technical complexity Integrations, framework adherence, traceability, data availability, COTS/legacy dominance, technical debt risk Security and compliance Ethical use guidelines, IP rights of code/ content, data privacy, architecture standards, legal, regulatory compliance Governance Transparency, explainability, accountability, monitoring, guardrails, avoidance of misuse and over-reliance Operating model Training for prompt engineering, process changes and conditioning for auto-generated code/content, refined roles and team structure
- 11. Our guiding principles 11 Keep the process and artifacts in sync Requirements to design to code to test Leverage LLMs for what they are good at Comprehend | Correlate | Code Harness contextual knowledge Onboard | Consume Value beyond productivity savings Speed of delivery | Improved quality
- 12. Knowledge driven software engineering 12 Knowledge onboarding The knowledge core Knowledge powered solutions Infuses enterprise-awareness Business processes, requirements/ user stories technology architecture, standards, compliance performance, security guidelines, operational data, procedure manuals and more… Harnesses contextual knowledge Industry knowledge - domain, standards, frameworks; TCS knowledge - learnings from various engagements, best practices across technologies and industries Leverages AI and generative AI UX design to front-end code generation, solution/technical design to backend code generation, functional test scenario test case creation and more…
- 13. New mobile app channel for a leading bank ‘ABC’ 13 Timeline Scope Technologies Inputs New channels and features, branding revamp Flutter(mobile), angular(web), java Business requirements, user stories, visual design, coding standards Rollout in <18 months
- 14. Reinventing SDLC in the age of AI 14 SDLC re-defined by leveraging contextual knowledge and orchestrating AI models responsibly along with humans Addressing the knowledge gap Closing skill gap Knowledge-driven engineering Onboard | orchestrate | consume Business process, systems and data Enterprise architecture principles and standards AI/ gen AI Define Design Compose/ integrate Validate Deploy Manage
- 15. Thank you Copyright © 2024 Tata Consultancy Services Limited
- 16. AI for Technology Modernization Ashvini Saxena 16
- 17. Technology Modernization Challenges 17 Modernization is not a one-time activity End up moving from one legacy to another legacy Very time consuming and high risk Most programs driven by Cost and Technology Risk; Fewer target Value generation
- 18. • Transform the existing applications to modern technologies faster & risk-free and future-proof by staying modern continuously by combining the power of Automation with human Augmentation using GenAI and a time-tested proven TCS’s application meta model • AI Models pre-trained with TCS’s strong domain & technology knowledge, 40+ years of experience in delivering large transformation programs and continuous learning of customer context • Bring more than 40-70% efficiency, 2x speed in delivering the modernization program with 100% accuracy on the outcomes Old/Existing Applications Meta Model/ Knowledge Repository Modern Applications Our GenAI Driven Technology Modernization Strategy 18 GenAI GenAI
- 19. Old / Existing application Target application List Employers List Employee Calculate PF List Employers List Employee Calculate PF Split Monolithic code to smaller, manageable services Generate functional specifications Generate target code components DevSecOps Deploy & Maintain Meta Model / Knowledge Repository Technology Agents Code Analyzers (COBOL, Java, VB, ABAB) Pseudocode Generator (Business logic, rules, workflow) Generation Agents Spec Generator (From doc, code, wireframe) Code Generator (Angular, Java, Python, .Net) Process Agents IPSafe/Guardrails (OSS, security) DevSecOps Insights (Dashboards) GenAI driven Technology Modernization – Demo 19 GenAI Assistant Agents (Virtual SME, Developer Assistant)