- 23–24 March 2027
- Amsterdam, Netherlands
Agenda
Business Intelligence and Analytics Summit 2027
Day 1: Tuesday 23 March 2027
–50 minutes
Registrations and Breakfast
–10 minutes
Opening Remarks
–30 minutes
AI Investment Decisions: Value, Cost and Board Expectations
Show moreShow less
- How the organisation evaluates and prioritises its data and AI investments.
- Criteria to determine whether an initiative receives further funding, remains in operation or is stopped.
- How recurring AI expenditure is identified, allocated and reviewed.
- What evidence senior management and the board require to assess performance and continued investment.
–30 minutes
Building Reliable Data Foundations for AI and Analytics
Show moreShow less
- The data engineering practices required to make enterprise data reliable for analytics, AI and agentic workloads.
- How ownership, data contracts, lineage, quality standards and pipeline monitoring work in production.
- Balancing shared data standards with domain ownership, while maintaining interoperability across platforms and AI systems.
–30 minutes
The Semantic Layer: Consistent Business Context for BI and AI
Show moreShow less
- How semantic models make business metrics and definitions reusable across BI, self-service analytics and AI applications.
- Where the semantic layer fits in modern data architecture and how it relates to governed metrics and AI retrieval.
- How organisations measure improvements in consistency, answer quality and decision-making.
- How disagreements over metrics and business definitions are resolved, and who is accountable for maintaining them.
–30 minutes
Coffee & Networking Break
–30 minutes
Panel discussionThe AI Stack Strategy: Concentration Risk, Procurement and European Alternatives
Show moreShow less
- Where dependency actually sits across cloud, models, platforms, data, identity and tooling.
- What European and open alternatives can realistically support today.
- The engineering, migration and operational costs involved in reducing dependency.
- How organisations assess portability, interoperability and vendor exit before signing long-term commitments.
–30 minutes
Enterprise Retrieval: RAG, Knowledge Graphs and Trusted Context
Show moreShow less
- When retrieval-augmented generation (RAG) is sufficient, and when graph-based or hybrid approaches may improve access to enterprise knowledge.
- How organisations handle entity resolution, permissions, changing source material and traceability.
- Comparing approaches on accuracy, latency, maintenance effort and operating cost.
–30 minutes
Data Products in Production: Ownership, Reuse, Quality and Lifecycle
Show moreShow less
- What makes a data product genuinely reusable across teams and business domains.
- Ownership models, service levels, quality expectations and discoverability.
- How adoption and reuse are measured after launch.
- When products should be improved, consolidated or retired.
–1 hour
Networking Lunch
–30 minutes
Token Economics and AI FinOps in Action
Show moreShow less
- How organisations measure the real cost of AI workloads beyond headline model pricing.
- Cost drivers across inference, retrieval, orchestration, retries, context and multi-step workflows.
- Forecasting demand and setting useful unit-economics measures for business owners.
- How FinOps changes architecture and model-routing decisions once workloads scale.
–30 minutes
Panel discussionThe EU AI Act in Practice: What Applies Now and What Must Be Ready by December 2027
Show moreShow less
- What organisations need to establish under the requirements already applicable, including relevant transparency and governance obligations.
- How to classify systems and understand provider and deployer responsibilities, with attention to the high-risk use cases relevant to the organisation.
- The controls, documentation and evidence needed for implementation, and how AI Act requirements fit with existing privacy, security and risk processes.
–30 minutes
Fireside chatThe 2027 Data & AI Operating Model: Decision Rights and Accountability
Show moreShow less
- How responsibility is divided between data and AI leadership, technology teams, risk functions and business units.
- Who owns data products, AI deployments, ongoing oversight and the business outcomes they are expected to deliver.
- Which capabilities organisations build internally, where they rely on partners, and how responsibilities change as systems reach production.
–10 minutes
Closing Remarks - Day 1
–1 hour 20 minutes
Cocktail Reception
Day 2: Wednesday 24 March 2027
–50 minutes
Registrations and Breakfast
–10 minutes
Opening Remarks
–30 minutes
AI and Agent Observability: Tracing Failures in Production
Show moreShow less
- What organisations actually need to monitor across models, agents, tools, workflows and downstream actions.
- How teams identify whether a failure originated in the model, the underlying data, an integration or the wider workflow.
- Identity, permissions and controls for non-human systems acting without a human session.
- Building audit trails that support investigation, governance and regulatory requirements without creating unnecessary monitoring overhead.
–30 minutes
Evaluating AI in Production: Reliability, Groundedness and Release Decisions
Show moreShow less
- How organisations test AI outputs against task-specific requirements instead of relying on a single, general measure of quality.
- How evaluation detects regressions as models, data and workflows change.
- How evaluation results inform deployment, restrictions, continued monitoring or withdrawal from service.
–30 minutes
Panel discussionGoverning AI Agents That Take Actions: Authority, Approval and Accountability
Show moreShow less
- How organisations classify agent actions by risk and decide which actions require human approval.
- How agent identity, permissions, overrides and escalation work in practice.
- Who is accountable when an agent takes an action, and how controls apply across multi-step workflows.
- How governance changes as agents move from recommendations to taking actions independently.
–30 minutes
Coffee & Networking Break
–30 minutes
Fireside chatWhen an AI System Must Be Stopped: Incident Response, Rollback and Recovery
Show moreShow less
- The signals that led to a decision to pause, roll back or withdraw a production AI system.
- The operational and business impact of unwinding workflows that other teams have already built around it.
- The controls, remediation and review processes introduced afterwards.
- What the organisation changed before allowing similar systems back into production.
–30 minutes
Conversational BI and Decision Intelligence: Where AI Improves Analytics
Show moreShow less
- How BI is evolving from dashboards and scheduled reporting toward conversational, embedded and AI-assisted analytics.
- Where natural-language analytics works today and where it still produces unreliable answers.
- How semantic models, governed metrics, and context affect the quality of AI-assisted analysis.
- What it takes to move from an answer or insight to a recommendation or business decision, and how to measure whether that helped.
–30 minutes
AI in Customer-Facing Operations
Show moreShow less
- How AI was deployed in service, sales or marketing, and why that workflow was chosen.
- What changed compared with the previous process, accounting for implementation costs and ongoing human effort.
- Where automation fell short and how the organisation decided when people should take over.
–1 hour
Networking Lunch
–30 minutes
Scaling Self-Service BI: Governance, Access and Adoption
Show moreShow less
- How organisations expand access to analytics while maintaining reliable metrics, appropriate permissions and clear accountability.
- How responsibilities are divided between central data teams and business users.
- How semantic layers, governance and AI-assisted analytics are changing the self-service model.
- Measuring adoption, trust and business value without recreating dashboard and reporting sprawl.
–30 minutes
Predictive and Prescriptive Analytics: Turning Forecasts Into Business Decisions
Show moreShow less
- Where forecasting, optimisation and scenario analysis are useful in business planning and operational decisions.
- How organisations move from predictions to recommended actions, and connect those recommendations to real workflows.
- What is needed to trust these models in practice, including data quality, model performance and the process for reviewing outcomes.
–30 minutes
Redesigning Work for AI: Roles, Skills and Accountability
Show moreShow less
- How AI changes responsibilities, workflows and the decisions employees and managers make.
- What capabilities employees need to use AI safely and effectively in the work they actually do.
- How organisations measure changes in productivity, quality or decision-making rather than relying on training completion or tool-usage figures.
–10 minutes
Closing Remarks
No sessions in this track on this day. Choose another day or show all tracks.
Join us in Amsterdam
Business Intelligence and Analytics Summit 2027
- Leadership, Value & AI Economics
- Modern Data Foundations
- Generative & Agentic AI
- Modern BI & Decision Intelligence
- Governance, Risk & Regulation
- AI Infrastructure & Sovereignty