Data without context is like sound without rhythm. It's just noise.
Just as rhythm gives music meaning, context gives data purpose. In the age of AI, context is the pulse that turns insight into understanding.
For brands striving to build deeper relationships with their customers, context is the bridge between data and empathy. It’s the essential ingredient that makes personalized, emotionally intelligent experiences possible.
We all envision a world of AI that’s built around trust, control and joy in every interaction. But this can only be realized when AI agents truly understand both the human and the product in every moment of engagement. That understanding doesn't happen by accident. It must be designed, engineered and continuously evolved.
Enter context engineering: the process of teaching machines how to listen, not just process. Teaching AI the who, when and why of every interaction.
Context engineering: connecting the dots that data can’t
Context engineering is an evolution of prompt engineering. In the same way that prompt engineering (done well) gives enough context to the LLM to enable accurate responses, context engineering extends the prompt by connecting dots across disparate data systems.
This could be systems that house customer and product data, for example a Customer Data Platform (CDP), Customer Relationship Management (CRM) platform, Content Management System (CMS) or a Product Information Management (PIM) platform.
“By engineering context into AI agent workflows, brands can deliver experiences that feel intuitive, responsive, and empathetically human.”
Mark Simpson
The role of customer context
Imagine a customer who’s navigating a complex buying decision. For example, they’re shopping for sustainably made hiking shoes suited for mountain terrain. Instead of clicking through endless filters, they open a chat with an AI agent and say:
"I'm training to climb a mountain and need shoes that are good for wet conditions but are also eco-friendly. I usually wear a UK 10 but sometimes a size up."
The agent, combining past purchase behavior, product specs, sustainability attributes, and data from the customer’s fitness tracker replies:
"Based on your training profile and past sizing, I’d recommend these three pairs of hiking boots. They are all made with recycled materials and have a strong grip for wet terrain. Two of them run small, so I've adjusted the size filter accordingly. Would you like to see waterproof-only options too?"
Not only has the agent provided the most personalized experience based on the context available, but the new parameters and preferences provided by the customer can now be remembered and applied across future interactions.
This isn’t just a transaction. It’s a conversation informed by context – which builds trust.
The role of product context
Equally important is understanding the product. CMSs, PIMs and other systems provide vital context such as what makes a product unique, how it measures up against others, whether it’s in stock, its sustainability credentials, or even user-generated content like reviews and styling tips.
Let’s say the customer wants to know whether a particular pair of boots will suit snowy conditions. An AI agent. empowered by rich product data, can respond:
“These boots are waterproof and have a high-traction sole, ideal for icy pavements. We also have a similar pair in stock with a faux fur lining if you want something warmer. Based on this product’s specifications, we can see that it meets your typical requirements for sustainability.”
When the context from both the customer and product is combined, the agent experience transforms from functional assistance to value-adding, delivering faster resolution, increased relevance, and deeper and more meaningful suggestions.
The hidden complexities of context engineering
Bringing these experiences to life is not without complexity. Context engineering involves more than just data integration. It requires careful orchestration of how, when, and where data is accessed and applied.
Some of the key technical challenges include:
- Latency and performance: Orchestrating and constructing multi-agent workflows which collect data from multiple systems in real-time (especially across microservices architectures) can introduce performance bottlenecks.
- Data freshness and accuracy: Ensuring that context is up-to-date, consistent across sources, and appropriately scoped.
- Privacy and consent: Respecting customer choices around data usage and ensuring that context-driven experiences comply with data protection regulations.
- Scalability: As the number of systems and signals grows, maintaining performance and reliability becomes more complex.
Solving these critical challenges requires careful design of systems and governing protocols.
How software vendors will become specialist data transactors
In this emerging agentic landscape, the role of the software vendor is being transformed like every other component. These platforms, which currently provide customer experience-enabled features, will instead become specialist data stores with agent interfaces, helping brands to curate and publish product data, content and information to a plethora of AI-infused endpoints. Many of these endpoints will be baked into:
- Devices that we use in our everyday lives today to engage with digital content (smartphones, tablets, in car entertainment systems).
- Devices that we use in our everyday lives, but not typically to consume digital content (home appliances, glasses, contact lenses).
- Devices that are yet to become mainstream (implants, haptic suits, VR/AR headsets).
- Devices that humans are yet to even imagine or conceive of, and are possibly invisible to the human senses (ambient AI, AI auras).
Curating the necessary content, activating the necessary context, and generally being able to orchestrate these experiences will be key. The technology vendors that provide these capabilities will be invaluable to the brand leaders of tomorrow, as much as they are today.
Successful vendors won’t be the ones who build the flashiest tools or the most features. They’ll be the ones who build interoperable platforms and robust ecosystems which allow AI agents to thrive across the experience design supply chain.
Systems integrators as experience architects
It’s not enough to just collect and store context. That context must be:
- Mapped across systems
- Cleaned for consistency and relevance
- Connected via APIs and middleware
- Activated in real-time
The role of experience architects will be orchestrating seamless flows of contextual data and empowering AI agents to deliver exceptional experiences across all channels. The vendors that can combine systems integration with AI-architected experiences will undoubtedly be the MVPs (most valuable partners) in an AI-powered world.
Context as an enabler of empathy
In the end, the goal of context engineering is simple: to turn data into connection. To enable AI to go beyond automation and become truly human-centered. For customers, that means relevance, ease and delight. For brands, it means loyalty, conversion and differentiation. And for AI agents, it means they can finally do what they were meant to do: serve. Because in an agentic world, context isn’t just king. It’s the connective tissue that makes empathy possible.
Contributor in this article
