Entity Resolution and Context Engineering on AWS
I've been obsessed with the role of context in decision-making systems for years. Back in 2005-2016 timeframe, while working at IBM as Chief Scientist of Context Computing, I used the puzzle metaphor to explain how context accumulates. I conducted experiments—mainly with kids—to understand the cognitive processes involved in putting pieces together. [Here's one of my Context Computing keynotes from an O'Reilly conference that brigs this to life.]
Fast forward to 2026: AI has changed everything. But one thing remains constant—context is still king.
Gartner captured this perfectly in their recent article "Context Engineering Is the New Prompt Engineering"
As they put it, "It's time to prioritize context over prompts with context-aware architectures, dynamic data and reimagined human-AI interfaces."
"Context engineering gives AI systems the situational awareness needed to act with relevance and precision. It enables AI to orchestrate actions, enforce policy, manage memory and tailor outputs ... all based on real-time knowledge, not just prompts."
At Senzing, our entity resolution SDK is both simple and powerful—it launches in 300-500ms yet scales to billions of records. The Senzing SDK functions as a composable AI agent that precisely identifies who is who and who is related to whom in your data. This rich context -- an entity-resolved knowledge graph (ERKG) -- is foundational to your AI workflows especially in your project is related to fraud, investigations, AML, KYC, MDM, or supply chain visibility and vetting.
Want to try it yourself?
Data scientists and engineers can test drive the Senzing SDK via AWS Marketplace. Spin it up quickly and see how easy it is to transform raw data into pristine entity-resolved knowledge graphs.
No one is more committed to democratizing entity resolution than Senzing. We deliver more capability and are innovating more quickly than anyone else in the ER space—consistently staying a generation (or two) ahead of the competition, and pulling further ahead every day.
Context is king. Start here: Senzing.
This is exactly the shift. Real-time, enterprise-wide context is what separates toy agents from systems you can actually operate. Curious how you’re thinking about freshness, ownership, and failure modes in the AWS setup — that’s where we’ve seen most complexity surface.
Kicking the tires on AWS sounds like a good way to test if your stack is ready for real workflows or just more demos. What's one blind spot most teams miss until it's too late Jeff?
You've certainly been ahead of the game in terms of driving context ahead of generating data for the sake of generating. It feels like today people are measuring parameters that a model can process rather than the quality i.e. situational awareness of what a model can produce
"Context Changes Everything" - quote from a business TV channel. Very, very true. Especially with the AI-driven flood of data that we're all swimming in. Without any context the data has much less value. That's a great keynote btw - always appreciate that message. Good stuff.