Evolving business intelligence with GenAI: the next step toward agentic analytics
By Andrew Berry, Senior Data Scientist, Bell Canada
Business intelligence (BI) is rapidly evolving. While traditional BI systems and dashboards for essential performance monitoring remain crucial, it is becoming equally important to enable analysts to leverage their domain expertise for deep strategic understanding, uncovering deeper insights – the "why" behind the "what.” However, this is often time consuming. Generative AI (GenAI) offers a way to accelerate this process, augmenting analysts' capabilities and simplifying the path to strategic insights.
It’s important to remember that GenAI isn't a magic bullet. To deliver reliable, advanced analytics, a successful implementation requires a deliberate strategy built on high-quality structured data, well-curated knowledge assets and precise business-specific rules.
Two key capabilities help GenAI enhance BI:
-
-
- Surface-level GenAI insights: providing quick answers to fundamental "what" questions (e.g., "What was the monthly revenue growth in Ontario?"), accelerating access to routine information. This is where we see immediate productivity gains.
- Agentic analytics: enabling automated, strategic analysis that tackles complex, multi-step "why" and "how" questions (e.g., "Why did sales spike last week at our Toronto store?"), driving deeper understanding and informed action. This is where GenAI truly transforms BI, moving beyond simple reporting to insightful analysis.
-
The BI maturity journey
To incorporate GenAI into their BI framework, organizations typically progress through the following stages:
Stage one, the dashboard era: heavy reliance on standard dashboards and reports. Analysts spend considerable time manually pulling data for ad-hoc requests. This stage is reactive; deeper insights take time, potentially delaying key decisions.
Stage two, getting smarter with GenAI Assist: GenAI handles surface-level questions, perhaps within existing BI tools. This accelerates routine queries, freeing analysts for more complex analysis. The "why" questions, however, still often require manual digging.
Stage three, the future is agentic (agentic analytics): Advanced AI systems proactively tackle complex questions, running analyses automatically or semi-automatically. This is transformative. Analysts become strategic advisors, guiding and validating the AI's findings. This stage requires well-managed knowledge assets, clean data, descriptive metadata, well-defined business rules and analytical playbooks.
The role of knowledge assets
Transitioning from Stage two to Stage three requires deliberately cultivating and managing knowledge assets. These assets guide AI systems to reliably and accurately perform complex analyses. Key components include:
-
-
- Structured data and metadata: clean, well-labeled data forms the foundation for accurate AI insights. Comprehensive contextual metadata is crucial.
- Domain expertise and rules: documenting how expert analysts interpret data, identify anomalies, and apply business logic is essential. Clear business logic ensures AI-driven insights remain relevant and actionable.
- Analytical playbooks: capturing common investigative sequences and workflows provides structure for AI systems to solve complex, multi-faceted questions, mimicking expert analytical processes.
- Iterative learning: feedback loops where user interactions and SME validation continuously refine the AI agents ensure accuracy and adaptation.
-
Advance your analytics and AI journey with Bell
Understanding the stages of BI maturity and strategically investing in knowledge assets is key to building a more impactful BI function. With Bell’s deep expertise in data strategy, AI implementation, and governance, we can help guide you through these stages, leveraging GenAI to unlock deeper insights and drive more informed decision-making. Let’s work together to elevate your Business Intelligence capabilities.