Real-time AI Experiences Can’t Advance Without a Universal Semantic Layer


With a universal semantic layer that organizes and standardizes your data, your company is ready to embrace AI wherever it leads.

Organizations today capture enormous amounts of data for improved operations, compliance, and analytics. In addition, they have decentralized to become faster and nimbler. This decentralization has resulted in increased data complexity, including multiple data silos and differing data definitions, formats, and datatypes. In this environment, it’s nearly impossible to produce consistent reports or business insights due to different data definitions originating from multiple source systems. It’s also impossible to fully embrace the power of artificial intelligence (AI).

The problem of data inconsistency is not new. Since the 1990s, data leaders have been talking about a single source of truth for their business intelligence and analytics. The quest for a single source of truth led to the creation of the “semantic layer” — a metadata and abstraction layer built on top of the source data that helps users autonomously access data using common business terms.

BusinessObjects built the first lightweight semantic layer into SAP BusinessObjects. Then MicroStrategy and others came along with their own semantic layers. All these early solutions suffered from the misconception that there would only ever be one tool to visualize and analyze the data.

Proliferation of semantic layers

The trouble with early business intelligence (BI) suites like BusinessObjects is that they are monolithic and not particularly user-friendly. Frustrated users adopted Tableau, Power BI, and Looker with their improved ease of use. Today, data analytics tools have grown and replicated across organizations, as have their many associated semantic layers, destroying any hope of a single source of truth.

Many BI tools, visualization tools, and many different, lightweight semantic layers create an enormous mess for the data team, who then plays whack-a-mole, constantly updating, correcting, editing, and adding to the business logic that’s sprinkled everywhere — a rough job for data engineers. But more concerning is the impact on the business: users get different answers, and trust erodes in the data itself, all because of inconsistency in metrics definitions and the lack of a single source of truth.

See also: The Semantic Layer’s Role in Analytics and Data Integration

The rise of the “universal” semantic layer

Separate from both the source and the output, a universal semantic layer organizes, simplifies, and accelerates the consumption of unified data to eliminate costly and redundant re-work across BI tools. But most importantly, it builds trust within the business with a consistent, standardized, and trusted single source of truth for data consumers, whether they are people or BI and AI systems.

What makes a semantic layer “universal?” A truly universal semantic layer should define all the metrics and metadata and deliver to all possible data experiences: BI software, data apps, large language models, embedded apps for customers, AI tools — wherever data is needed to serve a wide breadth of user experiences. In today’s world, there are more ways to deliver data than just a dashboard. A universal semantic layer ensures that data is consumable in multiple ways. It is also a prerequisite for powering the next generation of data- and AI-driven applications, accepting that there will be many different tools for visualizing and using that data and many different sources where data is stored.

See also: What’s the future of GenAI? Checking in with Industry Leaders

No AI without consistent data

The link between data and AI is unmistakable. High-quality data shapes AI systems into reliable interpreters capable of navigating and deriving meaningful insights from huge datasets. It’s not just the quantity but the qualityof data that pushes AI forward.

In other words, AI is forcing the universal semantic layer to be the center of gravity for enterprise data. AI tools require a semantic layer to understand the business context and definitions and avoid hallucinations. Once trained, AI can also improve the data models and definitions within the universal semantic layer by suggesting improvements to both the definitions and the code. AI agents backed by a semantic layer can further curate and democratize data by allowing business users to conduct natural language queries.

The rise of applied AI

AI is quickly spreading from the use of massive AI platforms such as ChatGPT and Meta AI to domain- and use-case-specific, applied AI apps — but the shift toward applied AI cannot happen without a universal semantic layer ensuring that data powering the apps is consistent, accurate, and optimized for performance.

As AI comes out of the lab and into the real world, the use of a universal semantic layer is essential. AI and LLMs are rapidly becoming the world’s biggest data consumers. Their efficacy hinges on the inputs they receive – specifically, consistent, relevant, and precise data. Without a single source of truth, AI systems cannot deliver accurate, trusted outcomes, and businesses can’t deliver the next generation of AI-assisted apps and experiences that will help them surpass competitors.

The benefits of an AI-ready universal semantic layer

A universal semantic layer that is AI-ready is needed to connect and work with diverse data platforms, protocols, and consumption tools. This decouples the data from consumption, thereby enabling the democratization of data analytics and AI in the enterprise.

For example, an AI-ready universal semantic layer can inform business users in real time and in context. Imagine that you are a sales operations professional who spends all day in an application like Salesforce. There is no time to learn a BI tool or jump out of Salesforce to do deeper analysis across sales, support, and purchasing data. Instead, a universal semantic layer makes it possible to embed AI-assisted analytics into a tool like Salesforce, allowing for the analysis to be done within the domain-specific business context — and via a near real-time process that can be as easy as querying an AI chatbot.

An AI-ready universal semantic layer can also power customer-facing applications that enable organizations to make the most of their data and their customer interactions. Imagine a financial services firm embedding an AI chatbot into investors’ monthly statements, allowing customers to query data about the growth of their investments across various sectors, timeframes, and investment types. Or, a skincare company embedding an AI-assisted recommendation app within their online storefront to make shopping more personalized based on skin types and tones. 

In the examples above, having an AI-ready universal semantic layer means that businesses can deliver AI- and data-powered products and services to employees and customers almost literally overnight. With a universal semantic layer that organizes and standardizes your data, your company is ready to embrace AI wherever it leads. 

Artyom Keydunov

About Artyom Keydunov

Artyom Keydunov is Co-founder & CEO of Cube, a venture-funded provider of a semantic layer for data apps. Prior to Cube, he co-founded Statsbot, a data platform.

Leave a Reply

Your email address will not be published. Required fields are marked *