Your Data Wants to Talk to You: The Age of Conversational Intelligence September, 2025 New York USA London UK Munich Germany Zug Switzerland Welcome Oleg Royz VP, Solution Consulting Oleg.Royz@dataart.com · 25+ years driving tech enabled growth through Data, Digital, and AI transformation. · Deep cross industry experience creating tangible results delivering ROI and Operational Efficiencies Alexey Utkin Head of Data & Analytics Lab Alexey.Utkin@dataart.com · 20+ years supporting clients' strategic transformation programmes with Technology and Data · Deep expertise in Financial Services, Fintech and Capital Markets industries Today's business environment is changing at a faster pace than ever before ...While having to navigate greater complexity than ever before... Macro / Geopolitical Uncertainty Less Predictability Fast Evolving Consumer Needs Supply Chain Disruptions Free Flow of Information Turnover / Knowledge Drain ... Moreover, The pace of AI adoption outstrips previous technological revolutions, forcing faster adaptation GenAI In this webinar we will address how to gain competitive advantage with conversational intelligence and what it takes create trusted and contextual data ecosystem Analysts and Dashboards Were Gateway to Business Intelligence Meet Sarah Director of Marketing She needed to understand which Q3 campaigns drove the most revenue in their Western region. Simple question, right? 1. Search through 47 different dashboards 2. Export data from three separate reports 3. Merge everything in Excel 4. Build pivot tables to analyze performance 5. Double-check her calculations Total time spent? Four hours. And she still wasn't fully confident in her numbers /4 Conversational Intelligence Addressing the gaps Now with AI/BI: Simply ask and get an accurate answer in seconds -- "How did our Q3 campaigns perform versus Q2?" -- AI-Bot: "Q3 conversions were down 10% compared to Q2. Would you like to see which channels were most affected?" -- "Yes, break it down by channel." It turns out email engagement as the weak spot. Another nudge leads to examine send times, revealing that early morning sends outperform others. With a few conversational turns, user pinpoints an actionable fix -- adjusting campaign timing -- without juggling multiple tools or manually correlating data. Conversational Intelligence: Redefining Business Intelligence Revolutionary Benefits Instant Data Conversations Talk directly to your data instead of building dashboards for every business question. Natural language queries deliver immediate insights. Curiosity-Driven Analytics Removes traditional barriers to data exploration. Business users can follow their intuition and ask follow-up questions without technical hurdles. Surface Hidden Needs Helps data teams discover real business requirements through natural conversation patterns and question flows. Operating Model Evolution 1 Governance & Guardrails New frameworks needed for conversational AI access controls, data quality validation, and automated compliance monitoring. 2 Hybrid Team Structure Data analysts evolve into conversation architects. Engineers focus on pipeline reliability. Business users become power questioners. 3 Self-Service Evolution Business teams will increasingly build custom agents and automated workflows, requiring new DataOps practices and citizen developer enablement. Strategic Implication: Conversational Intelligence doesn't replace your data team--it transforms them into enablers of business-driven analytics while requiring robust governance to maintain data integrity at scale. What it takes to get to Conversational Intelligence and AI/BI? What your data platform has Large Language Model Feedback Loops High Quality Data Knowledge Graphs Metadata Data Model LLM Data Glossary Semantic Model Evolving Context Usage Memory Access Controls Continuous Monitoring Features and Use Cases of Conversational Intelligence Data Conversation Tools · Focused on insight discovery, visualization and guided analysis · Raise business user data engagement · Suggests questions to ask, support curiosity · Grounded in specific data and semantic models · Embeddable into business and customer facing apps · Work with external LLM models and agents · Frameworks focus on end-to-end governance, semantic model curation, feedback cycles, tools and features for enterprise adoption and building accuracy and trust Data Reasoning Capabilities · Systems with broader AI reasoning that can synthesize across domains · More powerful, multi-hop reasoning, less grounded · Higher risk of hallucinations · Examples: Snowflake Intelligence, Databricks Genie Deep Research · Examples: Snowflake Cortex, Databricks Genie, Power BI Q&A / Copilot, AWS Q Business, ThoughtSpot Getting there - Pillars of Conversational Intelligence 1 2 3 Pillar 1 - The Data Bedrock Unified and Governed Data Ecosystem Pillar 2 - The Semantic Catalyst Semantic Models / Knowledge Graphs Pillar 3 - The Trust Engine Building Confidence via Feedback, Refinement and Scaling Pillar 1 - The Data Bedrock Analytics and Insights BI Dashboards Analytics Unified and Governed Data Ecosystem Data Apps Gen AI/BI Foundations Vector Stores Knowledge Graphs Semantic Model Semantic Search Knowledge Bases LLMS /GenAI Models Data Governance Data Catalog Metadata Management Data Lineage End-to-end Al-ready Data Quality Data Privacy Granular End-to-End Access Controls Data Marketplace Data Layer Data Ingestion Stream/Batch DWH, Data Lake, Lakehouse, Data Mesh Structured / Unstructured Data Data Federation / Virtualization Scalable Data Processing Data Sharing Pillar 2 - The Semantic Catalyst Building Semantic Model Data Dictionary - Definitions, Terms, Synonyms Knowledge Graphs, Ontologies, Entities, Relations Dimensions, Relations, Metrics, Measures, KPIs, Rules Domain Organization for Semantic Models Initial Adoption Partnership with Business Unit / Domain Gentle Adoption Cycle - User testing, validation, feedback Tools to support Semantic Model build out and KG mining Governance and Ops Model for AI/BI, Semantic Model, KG Pillar 3 - The Trust Engine Building Trust, Refinement, Scaling · Feedback cycle and continuous improvement. Learning from usage. · Reasoning / Analysis plan explanation · Verified Queries · Guardrails · Human (with the right skills) in the loop · Evaluations / Benchmarking · Monitoring and optimization (query complexity, token usage) 3 Flavours of Conversational Intelligence Capabilities Talk to your data Grounded, Governed, Focus on accuracy and consistency. Based on data catalogs, semantic/models Best in surfacing and working with existing organisational (historical) data. Reasoning, Research, Multi-hop Less grounded, includes "external documents", less consistent, more powerful in reasoning and interpretation. Harder to govern Agentic Conversational Intelligence More types of data, self-reflection/continuous learning, internal + external data and systems, complex reasoning, multimodal. Conversational Intelligence can use agents, as well as be a data/knowledge/insights tool for external agents via MCP/A2A The Future: Conversational Intelligence and Your Data in Agentic AI Beyond Conversation to Action AI agents that don't just answer questions but autonomously execute complex workflows based on data insights Preparing Your Foundation The pillars we've discussed become even more critical as the foundation for responsible, effective agentic systems "Organizations that invest in conversational intelligence today are positioning themselves for competitive advantage in the agentic AI future." Q&A Join the Conversation See What's Possible with \[Webinar Topic\] ­ Book an Expert Session! Contact us DataAnalytics.coordination@dataart.com Blending Data + Art to Achieve Real Business Results Getting real business value out of IT is equal parts science and art. DataArt combines the best of both worlds, offering world-class technical expertise ("Data") backed by a relentless drive to find creative solutions to our clients' most complex challenges ("Art"). True to our name, we fuse the technical and creative elements you need in a technology partner to be successful in a world defined by constant change. / 16