June 2025 Summaries

Sigma has introduced two enhancements to its Ask Sigma feature, providing admins with increased control and visibility over data and AI interactions. Users now have the flexibility to run Ask Sigma on AI models available within their existing data warehouse platforms like Snowflake and Databricks, enhancing security by eliminating the need for external models. Additionally, a new usage dashboard offers detailed insights into how Ask Sigma is utilized, enabling admins to track engagement, detect trends, and improve AI integration by understanding which questions are being asked frequently and by whom. This dashboard emphasizes transparency and security by keeping all usage data within the user's warehouse and allowing only authorized admins to access it. These updates aim to enhance governance, optimize data modeling, and ultimately foster a more informed and innovative use of AI in organizational decision-making.

Business intelligence (BI) has traditionally focused on analyzing past events, providing insights into what happened, but struggles when it comes to predicting future trends. This limitation has led to manual workarounds that are neither scalable nor reliable for forward-looking analysis. AutoML, or Automated Machine Learning, addresses this gap by automating the machine learning process, enabling BI teams to incorporate predictive analytics into their workflows without needing specialized expertise. AutoML streamlines data preparation, feature engineering, model selection, tuning, and validation, transforming the BI landscape by integrating prediction capabilities directly into everyday analytics tasks like querying and reporting. This allows teams to move beyond historical analysis, using predictions to anticipate customer churn, improve sales forecasting, detect fraud, and optimize workforce planning. While AutoML offers speed and accuracy by eliminating manual model-building processes, it still requires high-quality data input and careful oversight to avoid issues like overfitting. As AutoML and no-code machine learning tools become more prevalent in BI platforms, they empower teams to transition from descriptive to predictive and prescriptive analytics, expanding the scope of BI to include forecasting and strategic planning. This evolution signifies a shift in the role of BI professionals, who must now integrate traditional analytic skills with forward-thinking approaches to provide comprehensive data insights and recommendations.

AI systems, while often perceived as seamless and magical, are underpinned by a complex stack of layers that work cohesively to deliver precise and useful results. These layers include infrastructure, foundation models, orchestration and tooling, and the interface and applications layer, each specializing in different tasks to optimize AI performance. Infrastructure involves the physical and virtual components necessary for computing power, while foundation models are pre-trained systems that handle various functions. Orchestration and tooling ensure that AI responses are context-aware and stable, integrating prompts and memory components. The interface layer shapes user interaction, presenting AI's capabilities in an accessible manner. Governance runs throughout all layers, setting the rules for data access and AI behavior to ensure accountability and trust. Understanding this layered architecture allows data practitioners to better diagnose issues, improve systems, and foster trust in AI tools, highlighting the importance of clarity and modularity in AI development and deployment.

Server performance remains a critical factor in analytics, even when cloud platforms obscure it behind interfaces, as this abstraction can lead to blind spots that affect performance and trust in analytics tools. The management of servers, whether self-managed or vendor-managed, involves provisioning, scaling, and maintaining resources like memory and CPU, which directly influence the speed and reliability of dashboards and queries. Despite the promise of reduced hands-on management with cloud platforms, teams must remain vigilant, as vendor infrastructure decisions often align with broad usage patterns rather than specific needs, leading to performance bottlenecks. Reliable infrastructure is foundational for analytics adoption because users lose trust when tools are inconsistent, prompting them to rely on outdated methods. Observability tools can help teams identify and address infrastructure issues proactively, while asking the right questions of vendors regarding downtime, autoscaling, and support is crucial for accountability. Ongoing maintenance, including timely security patches and updates, is essential to prevent subtle degradations that erode system reliability over time, highlighting that ignoring infrastructure concerns can lead to costly consequences for analytics outcomes.

Spreadsheets, while indispensable for analyzing quantitative data, fall short when it comes to geospatial data, which involves information tied to specific locations and offers unique business insights. Geospatial data includes coordinates, addresses, and demographic information, revealing patterns and interactions that traditional data analysis cannot capture. Specialized visualization tools transform this complex data into actionable intelligence, enhancing decision-making in various business functions such as site selection, logistics, customer analytics, asset management, risk management, and sales territory optimization. Different types of geospatial visualizations, like point maps, cluster maps, flow maps, choropleth maps, and heat maps, each serve specific analytical purposes, revealing spatial patterns and relationships clearly and intuitively. However, challenges such as data cleanliness, appropriate map selection, performance issues with large datasets, and privacy concerns must be managed to ensure reliable insights. By integrating geospatial insights with existing business metrics and focusing on business questions, organizations can drive strategic decisions and gain a competitive edge over those relying solely on spreadsheets.

Despite significant investments in cloud data infrastructure, many companies still face challenges with inconsistent and unreliable data, often due to overlooked or poorly executed structured data transformations. This critical step involves cleaning, organizing, and standardizing data before it is used in reports and dashboards, ensuring that metrics are consistent and trustworthy. The process of structured transformation can prevent issues such as duplicated rows, schema mismatches, and null values that lead to misaligned metrics and eroded trust in data. Data Transformation Language (DTL) is highlighted as a scripting approach specifically designed for transforming raw data into analysis-ready formats, offering more precision and transparency compared to traditional SQL or GUI-based tools. By emphasizing structured transformation, businesses can improve the accuracy and reliability of business intelligence and analytics, ultimately fostering trust in their data insights.

The concept of a data lakehouse is emerging as a compelling solution for organizations facing challenges with traditional data architectures that separate data warehouses and data lakes. A lakehouse integrates the structured data management of warehouses with the flexibility of lakes, providing a unified storage solution that supports diverse analytics workloads without forcing teams to create isolated systems or add layers of complexity. This model allows businesses to maintain a single source of truth for data, streamlining operations and reducing redundancy, while also supporting scalability, governance, and performance tuning. Companies like Netflix, Shopify, and Databricks have embraced lakehouses to consolidate their data environments and enhance collaboration across teams. Despite the benefits, transitioning to a lakehouse involves complexities such as governance coordination, skill gaps, and performance optimization, which require careful planning and investment. Nonetheless, as organizations outgrow their legacy architectures, the lakehouse model is increasingly seen as an attractive option for achieving a balanced approach to modern data management.

Data errors in modern systems often manifest subtly, such as missing columns, unexpected values, or null entries, which can disrupt analytics and decision-making processes. These errors typically arise from overlooked assumptions or rapid changes in data sources and structures, leading to misalignment between expected and actual data inputs. Effective troubleshooting involves recognizing early warning signs, such as unusual dashboard behavior, and employing preventative measures like "canary" views and automated checks to catch discrepancies before stakeholders do. Collaborative efforts in debugging and understanding team behaviors can also help address silent data errors. Rather than seeking to eliminate all errors, the focus should be on learning from each incident to strengthen data systems and improve communication and documentation practices, ensuring that data integrity is maintained over time.

Effective feedback in analytics reporting is crucial for fostering clear communication, collaboration, and decision-making within teams. The process often faces challenges due to unstructured, delayed, or superficial feedback, which can lead to misunderstandings and misalignment on the intended purpose of reports. Constructive feedback is characterized by clarity, context, and collaboration, encouraging teams to ask questions, provide explanations, and document their insights. The ACED method, which stands for Acknowledge, Clarify, Explain, and Double-check, offers a structured approach to feedback, ensuring it is grounded, specific, and actionable. Analytics teams can enhance feedback by inviting input early in the development process, setting clear expectations, and ensuring psychological safety, thereby transforming feedback from a post-build critique to an integral part of the analytics workflow that improves the quality and utility of reports.

Sigma has been recognized for the first time in the 2025 Gartner Magic Quadrant for Analytics and Business Intelligence Platforms, an achievement attributed to its innovative approach to AI-powered analytics and data applications for over 1,400 customers. Sigma's cloud-native platform is noted for bridging the gap between data exploration and execution through features like governed AI, live collaboration, and real-time data interaction, which run entirely on the data warehouse. The company's future plans include developing AI assistants that facilitate intuitive data interaction and creating no-code tools for custom workflows, aiming to replace outdated BI systems with a more accessible, cloud-based solution. Sigma also maintains a strong partnership ecosystem and receives high customer satisfaction, reflected in a 4.8-star rating on Gartner Peer Insights, with plans to further showcase its vision and product roadmap in an upcoming webinar.

Misleading visuals in data presentations often arise not from intentional manipulation but from rushed decisions, default settings, or unchecked design habits, which can lead to confusion and mistrust. These visuals may break basic promises by showing incorrect data slices, visualizing information in a confusing manner, or implying conclusions not supported by the data, causing viewers to misinterpret the findings and potentially lose trust in the information presented. Issues like vague titles, improper axis scales, wrong chart types, and cluttered design can distort the intended message, while practices like cherry-picking timeframes and over-smoothing data can lead to misleading narratives. To maintain trust in visual analytics, it's crucial to pressure-test charts by ensuring clarity, relevance, and accuracy, and by questioning the visual choices made, to ensure they align with the true story the data is meant to convey.

Amid the growing complexity and volume of data, companies face challenges in providing quick, reliable insights without escalating costs and risks, which is where the ELT (Extract, Load, Transform) approach gains prominence. Unlike traditional ETL workflows that transform data before loading it into a warehouse, ELT loads raw data directly into a cloud data warehouse where transformations occur, leveraging the flexibility and scalability of cloud computing. This method preserves raw data, allowing for reprocessing if business rules change, and offers operational advantages by reducing bottlenecks and enabling complex transformations directly in the warehouse. While ELT facilitates agility and self-service BI by allowing analysts to work from a consistent raw data foundation, it also introduces challenges like potential data overload and the need for clear transformation governance to prevent inconsistencies. Ultimately, ELT encourages a strategic reevaluation of data workflows, advocating for automation where possible and intentional design in transformation processes, to enhance efficiency and support evolving analytics needs.

Snowflake Summit 2025, attended by 20,000 participants, was a bustling event characterized by numerous sessions, product launches, and discussions centered around AI and data solutions. Stipo Josipovic, a seasoned attendee, offered insights into the key themes and standout moments, highlighting the prominence of AI signage and Metadata's evolution. The NYSE demonstrated effective use of Sigma for data management, sparking interest in its pricing. Attendees were shifting from familiar tools like Power BI to Sigma, intrigued by its capabilities in handling data applications and AI, as exemplified by new features like Semantic Views and Cortex AISQL, which enhance data logic and SQL automation. Snowflake's keynote unveiled a series of launches aimed at creating a customer-centric, future-proof data foundation, emphasizing the importance of data movement and metadata in AI development. The significance of trust and transparency in AI, along with the rising importance of data movement and metadata, was evident, suggesting a shift in the data landscape where Sigma and Snowflake are becoming reference points for innovation.

Rows and columns form the fundamental structure of spreadsheets and databases, serving as the backbone for organizing and interpreting data. While they may seem simple, their arrangement is crucial for ensuring accurate data analysis and visualization. Rows typically represent individual records or observations, whereas columns define the attributes or variables of those records. Although spreadsheets and databases both use this tabular format, they differ in flexibility and structure enforcement, with spreadsheets offering more freedom but potentially introducing risks through formatting changes, and databases maintaining consistency through strict rules. The effectiveness of data analysis often hinges on the ability to reshape data structures, such as pivoting rows into columns, to fit the analysis context. Missteps in maintaining consistent data structure, such as merging cells or inconsistent data types, can lead to errors and inefficiencies. Advanced analytics tools help manage these structures by preserving column types and handling joins, ultimately ensuring that the underlying logic remains intact. Understanding and respecting the structure of data is essential for reliable analysis, as it impacts the ability to derive meaningful insights and trust the results.

Color plays a crucial role in data visualization, but its efficacy can be compromised when viewers have color vision deficiencies, which affects millions globally. Inclusive design in data visualization ensures that charts and dashboards remain accessible and effectively communicate insights, regardless of the viewer's ability to perceive colors accurately. The text highlights the importance of using colorblind-accessible palettes and incorporating design techniques such as direct labeling, pattern fills, and shape differentiation to enhance chart readability. It advocates for using tools like Coblis and Color Oracle to evaluate color choices, while emphasizing that accessibility is foundational, not optional, to effective communication. Additionally, it advises on choosing chart types that minimize reliance on color alone, such as bar and dot plots, and adapting less accessible formats like pie charts and heatmaps to maintain clarity. The text underscores that accessible design enhances communication, leading to faster decision-making and better outcomes, and it encourages a mindset shift towards inclusivity in data visualization rather than a complete overhaul.

Encryption at rest is a critical security measure that protects stored data by converting it into an unreadable format, ensuring that even if unauthorized access occurs, the data remains indecipherable without the correct decryption key. This concept is essential for data practitioners, who often encounter encryption in various analytics workflows, such as in data warehouses like Snowflake or Redshift, cloud object stores like Amazon S3 or Google Cloud Storage, and file systems holding exported data. While encryption at rest is often automatically applied by modern cloud platforms, understanding its implementation and key management practices is crucial, as improper configurations can expose data to risks, affect compliance with regulations such as HIPAA and GDPR, and influence the usability and trustworthiness of data. Unlike encryption in transit, which protects data as it moves across networks, encryption at rest focuses on securing data stored in various systems, adding a protective layer that complements other security measures. Awareness of where and how encryption at rest is applied can prevent potential vulnerabilities in analytics workflows, ensuring data safety and enabling teams to work confidently and efficiently.

Executives prioritize growth, risk management, and decision-making efficiency over technical analytics details, necessitating dashboards that offer clarity, context, and actionable insights rather than overwhelming data. Effective executive dashboards should focus on relevant metrics that directly impact strategic decisions, avoiding vanity metrics that don't inform real-world outcomes. Data storytelling enhances the dashboard's value by connecting disparate metrics into a coherent narrative, facilitating alignment and accountability across departments. Additionally, to support dynamic decision-making, dashboards should reflect real-time data and anticipate future trends rather than merely reporting past performance. Successful executive dashboards are organized around strategic themes, are simple, consistent, and allow for quick navigation to detailed insights when necessary. Building such dashboards requires a deep understanding of leadership priorities and an iterative process that incorporates feedback to ensure alignment with business goals, ultimately integrating analytics into the leadership rhythm and shaping organizational understanding.

Sigma has been recognized as the 2025 Databricks Business Intelligence Partner of the Year at the Databricks Data + AI Summit, highlighting its innovative contributions in the realm of BI through data apps, AI workflows, and live collaboration on governed data. Key announcements at the summit included Databricks' direct integration with SAP systems, eliminating silos and enabling real-time analytics, and the introduction of Mosaic AI, a platform for managing enterprise-grade AI agents with traceable and secure governance. Databricks SQL enhancements now offer serverless performance and full SQL support, which Sigma leverages to facilitate no-code workflows for business users. Sigma's integrations with Databricks also include the ability to run Python scripts within its interface, conversational analytics through Ask Sigma, consistent metrics management via Unity Catalog Metrics, and the real-time data capabilities of Lakebase integration. These developments aim to empower teams to work smarter, offering tools that blend operational efficiency with analytical depth, all on a foundation of governed and reliable data.

Sigma was named the Databricks 2025 Business Intelligence Partner of the Year at the Data + AI Summit, signifying the successful integration and practical use of Sigma with Databricks by customers. This collaboration enables business users to explore, analyze, and write back data directly on live lakehouse data without compromising data governance, access, integrity, or performance, thus resolving the typical tradeoff between user flexibility and data team control found in many BI tools. The partnership has led to reduced reliance on analytics teams for ad-hoc requests, the creation of customer-facing data solutions, and the operationalization of analytics beyond static dashboards. Sigma's CEO, Mike Palmer, emphasized this recognition as a validation of their vision for a new BI era, while Databricks' VP of Technology Partners, Roger Murff, highlighted the integration's capacity to deliver self-service analytics without sacrificing enterprise-level governance. At the summit, Sigma focused on engaging with customers about real-world applications of their technologies, underscoring the practical impact of their partnership with Databricks.

An operational data layer (ODL) serves as a crucial intermediary between back-end data systems and business-facing tools, ensuring data consistency and alignment across an organization. As data tools and definitions diverge, trust in data erodes, leading teams to create workarounds, often resulting in fragmented data interpretations and slowed decision-making processes. An ODL addresses these challenges by syncing and transforming data in real-time, allowing consistent logic to be applied across different departments, thus preventing the drift of data definitions. It not only supports analytics but also enhances operational efficiency by providing reliable, synchronized data that aligns with how teams work, fostering a unified, trustworthy system that minimizes the need for constant recalculations and verification. This integration reduces the time teams spend on managing discrepancies and increases their focus on deriving value from data. The implementation of an ODL is not about replacing existing systems but about enhancing them to maintain a cohesive and reliable data experience, ultimately supporting more effective decision-making and organizational momentum.

Early in their career, the author frequently said "yes" to tasks to learn and grow quickly, but at Sigma, they learned the value of saying "no" to focus on what truly matters. This approach helps in prioritizing high-impact work and making intentional decisions. When faced with a customer request to load audit log data directly into their warehouses, the team realized the complexities and potential issues, such as security concerns and maintenance burdens, and decided against it. Instead, they proposed an API solution for secure and reliable data access, reducing implementation complexity. The project emphasized the importance of collaboration, thoughtful decision-making, and the power of saying "no" to drive meaningful progress, exemplified by improvements in the audit pipeline and system resilience. The experience underscored that growth comes from prioritizing effectively and making strategic choices.

Composable analytics offers a flexible, modular approach to building analytics stacks, allowing teams to overcome the limitations of traditional, rigid business intelligence (BI) systems. Unlike monolithic setups that struggle to adapt to change, composable analytics enables the integration of specialized tools for different stages of data processing, such as ingestion, transformation, and visualization. This system is akin to constructing with LEGO blocks, where each tool performs a specific function and can be easily swapped or added as needs evolve. The approach encourages collaboration, iteration, and faster insights by allowing teams to work directly with data without being constrained by outdated workflows. Common myths about modular analytics, such as its presumed complexity or suitability only for technical teams, are debunked through advancements like no-code tools and improved orchestration platforms. This paradigm shift aligns closely with modern BI practices, emphasizing flexibility and adaptability over control, and is already embraced by many enterprises aiming to create sustainable, scalable data systems that evolve with their teams' workflows.

Data marts are experiencing a resurgence as organizations seek to balance the need for speed and autonomy with the requirement for consistency and control in data management. Unlike the rigid and silo-prone data marts of the early 2000s, modern data marts are lightweight, modular, and governed subsets of warehouse data tailored for specific domains, such as finance or marketing. They pull data from a central warehouse and deliver it in a structured format, offering a practical way to reduce complexity without requiring new platforms or teams. Data marts can be dependent, independent, or hybrid, each with varying levels of autonomy and governance, and they help streamline data access for business teams while reducing the central data team's workload. Their role is especially relevant as organizations move towards decentralization and data mesh strategies, allowing business units to operate more independently without losing central oversight. In the cloud era, platforms like Snowflake and dbt facilitate the creation and maintenance of data marts, providing modular architectures, declarative modeling, and user-friendly interfaces that support exploration while maintaining data integrity. The success of data marts depends on operational discipline, including clear ownership, feedback loops, and usage tracking to ensure they remain aligned, sustainable, and valuable.

Business intelligence (BI) often struggles with issues of trust and data integrity, prompting interest in distributed ledger technology (DLT) as a solution. Unlike traditional BI systems that rely on centralized databases, DLT offers a decentralized approach where data is recorded and verified across multiple sites, ensuring accuracy and transparency without needing a central authority. This model not only reduces the risk of data tampering and access bottlenecks but also allows for enhanced collaboration among teams by providing a common, verified data source. DLT's immutability feature ensures that once data is added, it cannot be altered without consensus, providing a reliable audit trail that is especially beneficial in industries with stringent compliance requirements. By embedding verification into the data structure and promoting shared truth, DLT can transform how organizations handle data governance, reducing the need for constant checks and enabling faster decision-making. While DLT is not a panacea, it offers a promising framework for BI teams seeking to enhance data trust and collaboration without the inefficiencies of traditional centralized systems.

Organizations often face challenges with data sourcing, which can lead to mistrust in analytics and decision-making inefficiencies. The issue arises when data is pulled from multiple, undocumented sources, leading to discrepancies in metrics and causing teams to duplicate dashboards or perform side calculations. This problem is compounded by the sheer volume of data available, necessitating a focus on structuring and clarifying data sources to ensure they accurately reflect business operations. Evaluating data sources involves assessing data quality, accessibility, scalability, and strategic relevance. Different teams prioritize different aspects of data sources based on their specific needs, with finance valuing stability, marketing prioritizing timeliness, product focusing on user behavior, and operations requiring speed and clarity. Data sources may need to be retired if they lack clear ownership, no longer align with business operations, create conflicting metrics, or require frequent manual intervention. Building a better sourcing strategy involves regularly reassessing data sources to ensure they support current business questions and removing those that no longer add value.

A data analytics audit serves as a comprehensive examination of the systems, reports, and processes that organizations rely on for decision-making, allowing them to identify gaps and misalignments in their analytics ecosystem. Unlike financial or IT audits, it reviews the entire analytics experience, from data collection to reporting, to ensure that insights are reliable and accurately interpreted. The audit process involves setting clear goals, understanding the data landscape, diagnosing quality issues, reviewing access controls, evaluating reporting tools, and mapping gaps and risks. It is crucial during times of organizational change to prevent deeper breakdowns and can be led internally or by external specialists. Effective audits prioritize transparency and continuous improvement, encouraging organizations to translate technical findings into business terms and make strategic adjustments that foster intentional decision-making. Ultimately, a successful audit not only highlights inefficiencies but also guides organizations towards maturity by fostering a culture of clarity, trust, and progressive change in data management practices.

Snowflake's Cortex AISQL revolutionizes the handling of unstructured data by enabling the processing of raw text, documents, and images directly using standard SQL, thus eliminating the need for manual intervention and overcoming the limitations traditional BI tools face with unstructured data. By integrating with Sigma, this technology allows teams to seamlessly incorporate unstructured data into their analytics workflows, enabling tasks such as classifying images or extracting terms from documents to be automated and scaled without increasing workforce demands. This innovation not only simplifies workflows across various industries, including legal and financial services, by allowing real-time analysis of documents but also ensures that decisions are based on comprehensive data rather than just easily analyzable information. The solution's ease of use, requiring no additional syntax or tools, further enhances its efficiency by allowing direct upload and analysis of files, thus integrating unstructured data with structured analytics in a governed and centralized manner.

Snowflake's introduction of Semantic Views represents a transformative development in data management by enabling metrics and logic to reside directly within the data warehouse, thereby eliminating redundancy and ensuring consistent data definitions across various consumption tools. This integration, supported by Sigma as a launch partner, addresses long-standing challenges of siloed logic and metric drift by allowing all downstream consumers, including BI tools and AI-powered interfaces, to access shared definitions of key business metrics. Semantic Views facilitate the calculation of complex metrics, such as aggregate-sensitive and expensive metrics, directly in the warehouse, thus ensuring accuracy and consistency across the data stack. As data warehouses evolve to support AI workloads, having semantics in the warehouse becomes increasingly crucial for generating reliable outputs from AI models. Sigma enhances this capability by providing a user-friendly interface to access and analyze Semantic Views, ensuring that updates in logic are consistently reflected across all tools, thus fostering alignment and trust in data across the organization.

The guide provides a curated roadmap for attendees of a summit focused on AI, data strategy, and business outcomes, offering a streamlined approach to navigating over 500 sessions and numerous exhibits. It emphasizes targeted sessions led by experienced teams, such as TradeStation's transition to self-service analytics and Saddle Creek's move from a legacy BI system to modern architecture with Databricks and Sigma, highlighting the efficiency gains and operational improvements achieved. The guide also features sessions on real-time analytics for responsible energy production, integrating analytics tools like Fivetran, dbt Cloud, and Sigma, and high-velocity decision-making in finance. Additionally, it includes networking opportunities such as an invite-only event for women in data and a rooftop party to foster community connections. The guide concludes by inviting attendees to visit Sigma's booth for demos and discussions on advanced analytics solutions.

At the Snowflake Summit 2025, Sigma was recognized for the third consecutive year as Snowflake’s Business Intelligence Data Cloud Product Partner of the Year, underscoring the successful collaboration between the two companies that serves over 1,000 joint customers, including notable names like DoorDash and Mindbody. Sigma stands out by allowing teams to work directly on the AI Data Cloud without the need for data extracts or separate systems, maintaining the same permissions and governance enforced by Snowflake. This seamless integration helps overcome common analytics challenges by enabling the creation of AI Apps that connect live data, centralizing analytics, AI, and decision-making processes. Sigma’s Generative Intelligence approach, which combines natural language processing, AI, and real-time cloud data, empowers users to build actionable data applications, distinguishing it from traditional BI tools. The partnership is celebrated for its consistency and excellence in working with the AI Data Cloud, as highlighted by executives from both Sigma and Snowflake. At the summit, Sigma is actively engaging with customers and hosting sessions to demonstrate the practical use of their platform in addressing last-mile analytics and utilizing Cortex AI.