If you are brave enough to drive in London, you will be all too familiar with the difficulty of navigating its complex web of narrow, winding roads, shaped by centuries of organic growth and redevelopment.
In comparison, New York, a city built with constraints and a clear plan from the outset, features a logical grid system, is predictable, and is designed for ease of movement. Navigation is straightforward.
Data architecture is no different. Some systems evolve organically, shaped by shifting needs and legacy decisions, while others are deliberately designed for clarity and scalability.
Some cities like Tokyo and Barcelona manage to weave centuries of tradition into modern, high-functioning infrastructure. It’s a patchwork of old and new. When structure is introduced gradually, it’s rarely simple, but it’s essential in making the city fit for the future.
Similarly, thoughtfully designed data architecture can streamline operations and improve client service.
A system that has grown organically can become difficult to navigate. Data becomes scattered, duplicated, and disconnected, turning what should be a strategic asset into an operational challenge. So before diving into AI, it is worth asking: is our data helping us move forward, or is it holding us back?
The Problem: Sprawling Data
The absence of a single holistic customer view has been a long-standing issue which should have been mitigated in the digital world. Instead, data is fragmented across multiple systems, often with duplicated records and inconsistent identifiers. Products are stored in isolated, non-integrated platforms with no direct connection between them. Rarely is there a process in place to ensure consistency across systems, let alone a unified view that reliably brings all this information together coherently.
For engineers and operations teams supporting these disconnected and disorganised systems, the task becomes akin to detective work, delaying decisions and hindering coordination. So, how can we shift from fragmented systems to a data architecture that enables faster, smarter decisions?
The Fix: Street Smart
Smart data begins with structure. A well-planned system employs consistent identifiers, such as a client ID, to connect customer data across platforms.
With this in place, answers arrive swiftly, without detours. It is possible to accelerate processes, but shortcuts require maintenance, expert knowledge, and constant updates. In many organisations, this kind of knowledge lives only in the heads of a few SMEs. When systems rely on this kind of deep, undocumented expertise, they become harder to maintain and even harder to modernise. The real question isn’t just about speed but whether the system is simpler to manage with one planned, consistent route.
Without standardisation, what seems to be a simple request can become frustratingly long-winded. Whether data resides in one system or many, consistency remains crucial. Clear labels, aligned formats, and reliable links make data easier to navigate.
Building for the Future: A Practical Framework for Smart Data Architecture
Just like a street network, data architecture shapes how smoothly everything moves. It serves as the blueprint for how data flows, is stored, accessed, and governed across an organisation, defining the structure of systems, the relationships between datasets, and the rules that ensure consistency, security, and scalability.
By setting out a practical framework firms can build a more structured, scalable architecture.
- Define
Defining your business objectives and use cases is a crucial starting point Ask what questions your data needs to answer:
- “Which products does my customer hold?”
- “What work is my team doing right now?”
This ensures your architecture is built around real business needs, not just technical ambition.
- Map
Next, map the current data landscape and take stock of what’s already in place. Consider where data is stored, how it is accessed, and what formats and identifiers are used.
- Establish
Create a canonical data model: a consistent structure for key entities such as clients, products, and transactions. Use standard identifiers like a client ID, and align formats and definitions across systems, to bring order to complexity.
- Design
Plan how data moves between systems. Use APIs, ETL pipelines, or event-driven flows. Decide where real-time processing is needed and where batch processing makes more sense. Ensure data lineage and traceability are built in.
- Implement (metadata and governance controls)
Metadata acts like signage, helping users understand what data means, where it came from, and how it can be used. Governance ensures data quality, access control, and compliance with regulations, building trust and scalability into infrastructure.
- Build
Design with adaptability in mind. Use modular architecture that supports new data sources, apply version control for schema changes, and maintain clear, accessible documentation. This will ensure data architecture can grow without descending into chaos.
- Validate
Road test your setup through real-world scenarios using actual business questions. Can you answer, “What products does my customer hold?” in seconds? Can AI tools like Copilot access and interpret the data reliably?
Building Roads to the Future
Just as cities balance heritage with modern infrastructure, organisations must evolve their data architecture with both purpose and foresight. When data flows freely, accurately, and consistently, teams can move faster, decisions become smarter, and innovation accelerates without getting lost in the backstreets of legacy systems.
Antonietta Price
Wealth Consultant
