AI Readiness Training Program – 5 Days (For Bank Employees, Kolkata)
Objective: Enable non-programmer employees (Excel-proficient) to use AI (cloud + local) for daily banking workflows, analytics, reporting & decision support with very less coding.
Tools used: Only open-source apps + publicly available data.
Day 1 – Foundations of AI & Prompting for Banking
Morning (3 hrs)
• Pre-assessment + ice-breaking + program overview
• Intro to AI & Generative AI → Impact on BFSI
• Banking use cases: customer support, fraud detection, credit scoring, compliance, reporting
• Prompt Engineering basics → structure of a good prompt, role/context/task, iterative prompting
• Examples: customer letters, RBI circular summarization
Post-Lunch (3 hrs)
• Zero-shot vs Few-shot prompting, structured outputs (tables, summaries)
• Risks: hallucinations, factual errors, privacy
• Demo: product comparison, MIS summaries
• Hands-on: draft FAQs, summarize branch reports, generate loan appraisal templates
Day 2 – AI without Internet: Local AI & Secure Practices
Morning (3 hrs)
• Local LLMs (Ollama, Llama, Deepseek) → need for confidentiality & comparison with cloud AI
• Demo: setting up Ollama
Post-Lunch (3 hrs)
• RAG demo → upload RBI circulars / SOPs → query in natural language
• Hands-on: load branch FAQs/policies → query for compliance answers
Day 3 – Data Analytics Beyond Excel
Morning (3 hrs)
• BI fundamentals: slicing, pivoting, filtering, drill-down, ranking, what-if, outlier detection
• Why Excel is limited for large data
• Intro to Apache Superset
Post-Lunch (3 hrs)
• Superset hands-on: import CSV → pivots, trends, drill-downs, ranking dashboards
Day 4 – Power BI for Dashboards
Morning (3 hrs)
• Power BI Desktop overview, components, architecture
• Connecting data sources, Query Editor → cleaning, merge/append, data modelling, measures
Post-Lunch (3 hrs)
• Visualizations: charts, tables, slicers, maps, cards, custom visuals, formatting
Day 5 – Power BI Q&A & Insight-Driven Dashboards
Morning (3 hrs)
• Power BI Service → dashboard vs report → publishing
• Quick Insights, filters: slicer/basic/advanced/top-N/measures/page/report/drill-through
Post-Lunch (3 hrs)
• Microsoft Copilot for Power BI → natural-language analysis & insight generation