Customer-based Context Engineering

I cannot stand a disconnected experience. I don’t think it’s a lot to expect when I’m a paying customer for a service that the company knows something about me and how I use their service, especially if I’m having trouble with it. But I’ve also been on the other side of that, trying to create that connection across a bunch of systems purpose-built to solve a specific problem for the business, but use customer information differently. And the bigger the company, the harder it is to figure this out.

Most companies build systems to reflect how they work internally. Over time, that creates a growing gap between how the business organizes information and how customers actually think about it. As systems multiply, customer data fragments across silos that make sense to the business, but not to the people using your services. The result is persistent friction that compounds over time.

That friction degrades customer experience and creates operational inefficiency throughout the organization. Every misalignment between systems and human thinking introduces new complexity, unnecessary work, and costly workarounds that scale with business growth.

This gap isn’t new, but as AI becomes central to business strategy, it becomes a bigger issue because AI amplifies both alignment and misalignment between systems and human context. Mid-market and growing SMB companies face particular vulnerability since they can’t afford wholesale system replacement yet still need to compete with more technically advanced players.

This is why I developed Customer-based Context Engineering. It is an approach to map, index, and expose customer context across all the different systems where customer information lives, creating coherence that isn’t usually possible without system replacement.

What is Customer-based Context Engineering?

The idea is simple: treat the customer as the primary organizing principle for how information is structured across the organization. That way, systems can work from a shared understanding of the customer’s perspective.

Customer Data Platforms (CDPs) and Customer Relationship Management systems (CRMs) handle some of those relationships and transactions effectively. However, they sit on top of business systems rather than inside them. Customer context engineering works at a deeper layer within the business. It links information about the customer across all systems (inventory, billing, ticketing/support, etc.).

Systems that share a customer context just work better. There’s no need to rebuild them from scratch; that’s too costly and risky. Instead, we create a smart index that connects customer information across systems. This index knows how to find and interpret it no matter where it lives.

Consider how this might work in practice: When a customer calls support about an issue with their service, the agent doesn’t need to navigate through multiple systems or ask them to repeat information the company already has. Instead, the customer context index immediately provides the agent with a complete view of the customer’s products, recent interactions, billing status, and service history. All of the customer’s data is pulled from different backend systems but presented as a coherent whole. The agent can immediately focus on solving the problem rather than hunting for information across disconnected databases. This same contextual awareness can extend to self-service interfaces too, giving customers direct access to relevant information without needing to understand your internal system boundaries.

Real-World Impact

When Charter, Time Warner Cable, and Brighthouse merged in 2016, a new entity, Spectrum Enterprise, was formed. As part of that, we inherited more than 40 disconnected portals operating with independent logic. The ask was to create a self-service portal for Spectrum Enterprise customers. However, to do that, we needed a consistent way to identify our customers and their data across a recently combined back-office from the merger.

What began as a customer experience initiative quickly revealed itself as a fundamental systems architecture challenge. The underlying problem extended beyond fragmented websites and interfaces. Most of the technical ecosystem didn’t understand or reflect how clients naturally perceived their relationship with Spectrum Enterprise.

The team came up with a novel idea. They decided to use big data and machine learning to process data from all business sources (millions of records) to algorithmically understand, map, and create an index that would function as a key to all the relevant information a client would expect to see or interact with on a self-service portal.

The results were substantial: over 60,000 clients adopted the new portal, support calls decreased by 60%, and the project created a unified customer dataset that worked across all channels while preserving the core systems the business relied on for daily operations.

This approach is not rocket science, but it requires thinking differently about your business data and systematically building a way to understand your customers.

How to Find Customer Context

Over time, I’ve identified five core disciplines to help find and codify customer context, each addressing a different layer of system behavior affecting how it is captured, surfaced, and used throughout the organization’s technical ecosystem.

1. Service Blueprinting with Data Anchors

Service blueprinting with data anchors connects customer experiences to their underlying data sources across business systems.

Traditional service blueprinting maps customer touchpoints to the people, processes, and tools involved in each interaction. Data anchors extend this approach by explicitly identifying where critical customer information lives at each touchpoint, not just which system contains it, but the specific data structures, fields, and relationships that support the experience.

For example, when a customer views their order status, data anchors map this interaction to the order management database, the specific tables containing status codes, the customer identification systems, and any transformation logic that converts internal codes into customer-friendly language. This detailed mapping reveals where customer data sits across all business systems, creating the inputs to a customer data model.

2. Customer Data Modeling

Customer data modeling creates structured representations of customer information that mirror how customers naturally think about their relationship with your business while working within unavoidable technical constraints.

Conventional data modeling starts with system capabilities and technical constraints, building outward from what’spossible rather than what’s needed. Customer context engineering inverts this approach, beginning with customers’mental models and working backward to align technical implementations with the way they naturally think about their relationship with the business.

Consider a financial services company where investment products, checking accounts, and mortgage loans exist in separate systems. While each system efficiently manages its specific function, customers think of their relationship holistically. A customer data model would create connections that recognize when “Sarah’s college fund” relates to both an investment account and a recurring transfer from checking, enabling advisors to discuss financial goals rather than isolated product details.

3. Data Interrogation & Cleanup

Identify and clean up data that is missing, incomplete, or incorrect in context of the customer data model.

This step involves systematically examining your existing data to identify quality issues that would undermine your customer context efforts. Common problems include missing contact information, outdated preferences, duplicate records, and inconsistent formatting across systems. By thoroughly assessing data quality before building new experiences, you ensure that the contextual layer you create will provide reliable information.

4. Indexed Data Layer

Build bridges between systems without replacing them, creating an indexed layer that translates customer language into system logic without disrupting existing operations.

This approach creates a thin layer that maps relationships across systems while allowing those to continue functioning without major changes, instead of consolidating everything into a single database.

5. API & AI Agent Interface Enablement

Create simple access points that let both AI and humans use your customer context effectively.

Developing these endpoints ensures the same customer understanding is available to support agents, AI systems, self-service interfaces, and internal tools. This makes customer context easy to take advantage of both inside and outside of a business.

Each of these steps has value to a business on its own, but going through all five provides a system to easily layer customer context into most any business.

Why Leaders Should Care

Most companies treat customer experience, data architecture, and operational efficiency as separate domains with separate owners. But when you examine them through a systems lens, you see they’re intrinsically connected around the customer. When technical systems can reflect how people naturally think and behave, customer context can be a competitive differentiator. Support becomes more efficient. Sales becomes more effective. The entire business runs with less friction and greater resilience.

Customer-based context engineering reduces the risk of AI initiatives failing due to incomplete or misaligned data. This is a common problem when organizations attempt to implement AI on top of fragmented sources that don’t provide the context needed for meaningful interactions.

For product leadership, this approach can transform how customer value is measured and tracked. When systems share a unified customer context, product teams can connect technical metrics to actual customer outcomes across the entire journey. Instead of measuring success through proxy metrics within isolated systems, leaders can track how specific investments directly impact customer lifetime value, retention, and satisfaction. This creates deeper visibility into which initiatives truly move the needle for customers, enabling prioritization and clearer connection to business results for product decisions. Product roadmaps become living documents that show the way to measurable customer value delivery.

The strategic benefits compound over time. As customer context becomes embedded in your systems, each new feature or service inherits this contextual awareness automatically. This creates a virtuous cycle where operational improvements directly enhance customer experience, which drives adoption and reduces support costs. Organizations that implement this approach can maintain personalized, high-touch experiences as they scale, breaking the traditional tradeoff between growth and customer intimacy.

Getting Started

Customer context work is most successful when led by cross-functional teams with executive support, requiring collaboration between technology, product, and customer experience leadership to ensure alignment across organizational boundaries.

Start with a touchpoint that causes significant friction with your customers (either anecdotally or through customer support metrics). Then, walk through the first couple of steps above. You’ll find there is just as much value in the conversation to get to a unified customer context as there is in implementing it in the system.

I’ve seen how aligning systems to human context creates both good design and good infrastructure. As both a frustrated customer and a product leader, I know that companies who recognize this connection gain advantages that go beyond individual products or features, creating value that compounds over time.