Talks in English

“From Vision to Implementation: Finding AI Use Cases That Actually Work”

Duration: 45 minutes | Format: Keynote + Interactive Session

Description:
Most AI projects don’t fail because of technology – they fail because of poorly defined use cases. In this talk, I present the proven Pandata AI Innovation Canvas – a framework developed from over 50 projects. We’ll walk through the entire process together, from problem definition and data source validation to feasibility assessment.
Attendees will learn:

• How to distinguish real business problems from “AI hype”
• The 8-step Canvas for systematic use case development
• Which questions must be answered before any AI project
• Practical red flags: When you should NOT start with AI
  Target audience: C-Level, Product Managers, Innovation Teams
  Takeaway: Canvas process for your own use case exploration

“AI-Ready in 90 Days: Building the Data Foundation for Successful AI Projects”

Duration: 40 minutes | Format: Technical Deep-Dive

Description:
“We don’t have data” is the most common excuse for failed AI initiatives. The truth: Most companies have data – it’s just not accessible, structured, or documented. This talk shows the pragmatic path from chaotic data silos to an AI-ready infrastructure – without years-long data warehouse migrations.
Attendees will learn:

• The 4-layer model: Raw → Clean → Object → Reporting
• Use-Case-Driven Development instead of Big Bang Approach
• How to become productive in days rather than months with Query Pilot
• Vendor-agnostic architecture patterns for RAG & Agentic AI
  Target audience: CDOs, CTOs
  Takeaway: “AI Data Readiness Assessment” checklist + Architecture Blueprint

“Rethinking Efficiency: Process Optimization with AI – Done Right”

Duration: 35 minutes | Format: Case Study Showcase

Description:
AI for process optimization sounds appealing – but when does “Make” make sense and when should you “Buy”? Using concrete examples from our portfolio (support automation with RAG, budget allocation, pricing optimization), I show how to maximize ROI and avoid common pitfalls. Spoiler: Not every process needs a custom LLM.
Attendees will learn:

• The Make-vs-Buy matrix for AI solutions
• When Agentic AI makes sense (and when it doesn’t)
• How to identify and scale quick wins
• Realistic success metrics beyond “AI-washing”
  Target audience: CPOs, CSOs, Department Heads, Process Owners
  Takeaway: Decision Framework “AI Process Optimization Canvas”

“Chat with Your Data Warehouse: LLMs as the New BI Interface”

Duration: 30 minutes | Format: Live Demo + Technical Talk

Description:
SQL knowledge as a prerequisite for data analysis? That’s yesterday’s thinking. With Query Pilot, I demonstrate live how language models become a natural interface to your data warehouse – 100% privacy-compliant through exclusive use of metadata. A game-changer for data democratization.
Attendees will learn:

• How text-to-SQL works in practice (including limitations)
• RAG architecture for structured data
• Privacy-by-Design: Metadata vs. actual data
• Integration into existing BI landscapes
  Target audience: Data Leads, BI Teams, CDOs
  Takeaway: Live demo of Query Pilot + Technical Architecture Guide