While much of the AI industry has been preoccupied with Unstructured Data, Engine AI was built from day one to work across both Structured and Unstructured Data. 

Enterprise AI adoption is accelerating. As firms move beyond experimentation, one point is becoming harder to ignore: trustworthy Enterprise AI has to be grounded in Structured Data. The databases, data warehouses, and proprietary models that organisations have spent decades building. 

In Capital Markets, this has never been up for debate. The decisions that matter are built on Structured Data. It is how Engine AI was designed from the outset. 

The rest of the market is now catching up. At Nvidia's GTC, Jensen Huang, founder and CEO of Nvidia, put it clearly: 

"Structured data is the foundation of trustworthy AI." 

For us at Engine AI it reflects three principles we have built around from the start: 

1. Structured Data is the foundation.  

2. Domain Knowledge is the differentiator.  

3. Intelligence has to be embedded into the way people work, not bolted on as a separate tool. 

Structured Data Is the Foundation 

Jensen's keynote at GTC was not primarily about data. It was about the future of AI infrastructure, inference at scale, and the next generation of computing systems. 

But woven through it was an argument that deserves more attention than it received. 

The shift from consumer AI to enterprise AI is not just a market transition. It is an architectural one. Enterprises have spent decades organising their operations around Structured Data. Databases, data warehouses, proprietary data models, risk systems, pricing engines. That infrastructure represents years of investment and institutional knowledge. Trustworthy AI, Jensen argued, has to be grounded in it. 

Engine AI was built on this conviction from the start. We began in Capital Markets, working with enterprises whose world has always been both: Structured Data such as prices, estimates, and portfolio positions, and Unstructured Data like research reports, analyst notes, and earnings transcripts. 

The firms we work with invested decades and significant capital into that combined data infrastructure. Our job is to connect it, contextualise it, and deliver it as intelligence. That means applying domain-tuned analytics and ML models that transform raw data into accurate, explainable insight. Not just aggregating sources. Genuinely understanding what the data means in a Capital Markets context. 

While the broader AI industry focused on Unstructured Data, Engine AI was designed from day one to work across both Structured and Unstructured Data. The industry is only now arriving at the same conclusion. 

Domain First: Depth Beats Breadth in Capital Markets 

Jensen made another argument at GTC that deserves equal attention. 

As general-purpose AI commoditises, the companies that build a lasting moat will be the ones that go deep into a specific domain, understand its logic, and build AI that is genuinely useful within it. 

Domain Knowledge is not a feature. It is the architecture. 

In Capital Markets, depth is what makes AI trusted enough to be embedded in the decisions that matter. 

Capital Markets has its own data relationships, its own analytical logic, its own language. A research analyst at a bank does not need a general-purpose assistant. They need intelligence that understands coverage universes, earnings revisions, and macro signals. And how those connect to the decisions they make in their specific role. 

A trader needs something different. So does a wealth advisor, a risk manager, and a product manager at a data company. 

This is why Engine AI delivers intelligence personalised across four levels. 

1. Industry level – a Capital Markets knowledge graph and ontology connects and contextualises data across Structured and Unstructured sources simultaneously.  

2. Firm level – proprietary data, models, and business rules are embedded directly into the Engine, so every output reflects the logic and context of that specific organisation.  

3. Team level – persona-specific workflows are designed around how each role operates.  

4. User level – workflows adapt over time so intelligence becomes more relevant with every interaction. 

Without domain expertise as the foundation, none of that personalisation is possible. You cannot contextualise data you do not understand. You cannot build firm-specific logic without knowing what that logic should look like. 

But knowing what matters is only half the equation. How that intelligence is surfaced, and to whom, is what determines whether it gets used at all. 

Intelligence Has to Be Discoverable 

Nick Turley, Head of Product at ChatGPT, was asked recently how OpenAI plans to reach the next billion users. 

His answer had nothing to do with model capability. 

The problem, he said, is that ChatGPT still looks too much like a computer terminal. Delegation is not a natural skill. Most people do not know what the tool is capable of or where to start. 

The solution is discoverability. Interfaces that guide users toward value rather than requiring them to already know how to extract it. 

From day one, Engine AI was designed around this principle. The interface is not a wrapper around the intelligence. It is part of the intelligence. 

Capital Markets professionals should not need to become prompt engineers to get value from AI. The right insight should be there when they need it, in the context they are already working in. 

What This Looks Like in Practice 

Every Engine AI interface is built from the ground up around how a specific role thinks, works, and makes decisions. 

The intelligence does not wait to be asked for. It is organised, surfaced, and delivered around the workflow before the user even formulates the question. 

And for the intelligence that needs to reach people inside the systems they already work in, Agents, APIs, and MCPs ensure it arrives there directly, in the flow of work, without friction. 

In Trading Intelligence, institutional-grade data across Equities, ETFs, FX, and Crypto is distilled into explainable insights across Fundamentals, Technicals, Flows, and Sentiment. Structured around how traders monitor markets. No query construction required. 

In Research Intelligence, internal research and external data are unified in a single environment. Analysts move fluidly between Structured Data exploration and natural language queries. Relevant signals surface proactively. And through Vibe Analytics, users can customise their own workflows using natural language, without engineering support. 

Capital Markets Intelligence, built in partnership with KX, brings Structured and Unstructured Data together in a single environment. Combining KX's data infrastructure with Engine AI's Domain Knowledge layer and Intelligence Interface, it helps data and technology companies unlock new revenue streams and drive deeper adoption with existing clients. 

Timing is Everything 

Three years ago, we were early. 

Today, the market has arrived at the same thesis Engine AI was built on. Structured Data as the foundation. Domain Knowledge as the differentiator. Intelligence embedded into the way people work. 

The conversation has caught up. The question now is who acts on it first.