Talk to Your Portfolio: How AI + Data Object Graphs Are Transforming Investment Intelligence

What if you could talk to your investment portfolioโ€”ask complex questions in plain English and instantly get nuanced, actionable answers?

At Dataception Ltd, we're making that a reality.

We're combining AI with Data Object Graphs (DOGs) to create a fundamentally new way of performing investment analysisโ€”one that's dynamic, intuitive, and powerful enough to support the demands of todayโ€™s asset managers.


From Class Models to Data Object Graphs

Traditionally, portfolios have been modelled using flat tables, static objects, or at best, traditional knowledge graphs. These approaches struggle to capture the living, breathing nature of modern investment ecosystems.

So we reimagined the portfolioโ€”not as a dataset, but as a graph of interconnected entities with built-in behaviour.

We started with:

  • A class model: portfolios, assets, transactionsโ€”complete with operations as methods (like pricing algorithms, ESG scoring, risk models)

  • Then instantiated it with real-world data: for example, Johnโ€™s portfolio with Apple shares and renewable energy bonds

The result is a Data Object Graph (DOG)โ€”a hybrid graph where:

  • Nodes represent both data and executable logic

  • Relationships are first-class citizens

  • Traversals dynamically execute operations along the way

This allows us to simulate and interrogate real-world scenariosโ€”not just describe them.


Ask Questions in Plain English

The real breakthrough came when we connected a small, local LLM (language model) to the DOGโ€”enabling natural language queries to be translated into graph queries using GQL, the graph query language standard.

Some real-world examples:

๐ŸŸข "What is our exposure to oil & gas companies?"
โ†’ The LLM translates the question
โ†’ The graph executes a traversal, with nodes calculating exposure as we go
โ†’ Result: 3.4% exposure across 7 positions

๐ŸŸข "Show all investments with AAA ESG ratings in renewable energy."
โ†’ The graph filters and returns all matching assets
โ†’ Immediate insights for aligning with sustainability goals

๐ŸŸข "What would happen if we divested from all companies with poor environmental ratings?"
โ†’ The LLM generates a multi-stage graph traversal
โ†’ Execution nodes simulate rebalancing logic across the graph
โ†’ The system returns impact analysis and potential reinvestment scenarios

All of this happens on-premiseโ€”LLM and graph engine includedโ€”keeping sensitive financial data secure within your infrastructure.


The Benefits: From Insights to Impact

This approach is more than a tech upgrade. Itโ€™s a transformation in how investment professionals interact with data:

  1. Model Reality More Accurately 
    DOGs capture the actual structure of financial systemsโ€”not just simplified abstractions

  2. Democratise Access to Insight 
    Portfolio managers, analysts, even ESG officers can ask deep questionsโ€”without writing code

  3. Accelerate Innovation 
    Add new asset types, sustainability metrics, or regulatory frameworksโ€”without breaking your system

  4. Embed ESG into Core Workflows 
    No bolt-ons. ESG and financial logic live together inside the same graph

  5. Stay Local and Secure 
    With everything running inside your firewall, compliance teams sleep easy


DOGs + AI: A New Era for Investment Analysis

This hybrid graph model enables a new style of thinkingโ€”one that reflects how investors and analysts actually reason about portfolios. Not linearly. Not in rows and columns. But as dynamic networks of interdependent elements that shift and recalculate as new data emerges.

And when combined with AI, these systems become not just queryable, but conversational.

We're moving from:

  • SQL queries โ†’ natural questions

  • Static dashboards โ†’ living models

  • One-size-fits-all analytics โ†’ personalised, context-aware insight

At Dataception, we call this talking to your portfolioโ€”and it's changing how finance thinks about data.



With Dataception Ltdโ€™s Data Object Graphs, AI-powered finance is just a walk in the park. ๐Ÿพ