Kyvos Provides Semantic Context for BI and AI

I have recently written about the renewed focus on semantic modeling and semantic layer software and the role they play in providing context for generative artificial intelligence (GenAI) and agentic AI. Semantic modeling has always been a critical enabler for business intelligence (BI), adding agreed-upon definitions and meaning to data to facilitate business understanding. It is also now proving essential as a facilitator for agentic AI initiatives involving automated decision-making based on multistep reasoning. This should be good news for providers of semantic layer software, such as Kyvos, which have expanded their offerings with additional capabilities to support the specific requirements of enterprise AI.

Kyvos Insights was founded in 2015 to create a semantic layer to accelerate analysis of data using BI and visualization tools. The concept of the semantic layer can be traced back to the early 1990s and had previously been implemented by multiple BI tool providers to accelerate multidimensional analysis of data through a combination of preaggregation and semantic modeling. Kyvos was part of the next wave of semantic layer providers that delivered standalone semantic layers that could be used to accelerate analysis of data in cloud data warehouses and data lakes using multiple BI and visualization tools. Kyvos was rated Exemplary and a Leader in the 2026 ISG Buyers Guide for Analytics Emerging Providers. While acceleration of BI continues to be a primary use case for semantic layers, recent attention on semantic layers has been driven by the need to provide business context for AI agents that enable automated decision-making. I assert that through 2028, analytics and data providers will prioritize the development of semantic modeling capabilities to facilitate the execution of autonomous AI agents.

Semantic layers provide two primary benefits. One is preaggregation of queries in online analytical processing (OLAP) cubes to accelerate analysis. While important, this has always been a relatively niche activity that is used only for the most demanding use cases. The other primary benefit is the combination of semantic definitions and semantic modeling, which provides conceptual information that reflects the meaning of entities, facts and relationships and establishes agreed-upon definitions related to business logic that enable users to analyze data without the need to become experts in the structure of the data or the SQL query language.

Semantic modeling is more broadly applicable to a wider range of use cases and remains a significant business challenge. More than one-half (52%) of participants in ISG’s Data and AI Study stated that a lack of access to semantic logic is a common impediment to data initiatives, making it hard to understand what the data means. Despite its importance, enterprise spending on semantic modeling is typically lower than on other data and analytics tools. Only 9% of participants in ISG’s Data and AI Market Lens Study named semantic data modeling as one of their five largest funded data activities in 2025, compared to reporting and dashboards (56%), data integration (53%), data quality (43%) and data discovery (27%).

This indicates pent-up demand for the value of semantic capabilities that is not being served by current levels of investment. Indeed, the level of investment in semantic layers and semantic modeling can be expected to increase given its role in providing business context for AI agents. As I previously explained, access to semantic definitions is required to provide a foundation of business context that facilitates search-based data discovery via knowledge graphs, as well as natural language analytics interfaces and agent-based decision-making. To make context-aware decisions and take actions based on perception and reasoning, agents require access to semantic definitions as part of an evolving context stack that also involves a clear understanding of business goals and KPIs, organizational structures and relationships, actions and outcomes, skills and tools, regulatory requirements and policies, as well as planning cycles and strategies.

Kyvos has evolved its product offering with a view to providing complete business context for AI that addresses semantic definitions, metrics and dimensions, along with business entity relationships, policies and rules, as well as an agentic workspace that enables users to assemble and govern context and tools required for AI use cases. Kyvos also offers a business analytics agent, called Dialogs, that automatically answers natural language questions with responses that are rooted in semantic definitions, as well as business policies and reasoning and planning. Dialogs provides integration with multiple AI models and can be embedded in productivity tools, including Microsoft’s Teams and Salesforce’s Slack. The primary claimed benefits of this combination of functionality include significant improvements in AI accuracy, query performance and token efficiency. Kyvos’ semantic foundation can also continue to be used to address its more traditional use cases of providing semantic consistency, advanced data modeling and acceleration for BI tools including Salesforce’s Tableau, Microsoft’s Power BI and Excel and Google’s Looker.

Many providers of data platforms and analytics tools are now delivering semantic modeling capabilities at the data layer rather than the analytics layer, driven by the need to provide context for AI. This poses a new competitive challenge for semantic layer specialists such as Kyvos. However, it may also drive increased customer demand from enterprises that are cautious about becoming overly dependent on a single data stack provider or facing the prospect of trying to integrate multiple semantic layers from multiple providers. I recommend that enterprises assessing providers capable of delivering semantic and context layers for AI and/or BI include specialist semantic layer providers, including Kyvos, in their evaluations.

Regards,

Matt Aslett