In old-school strategy games like Civilization 5, your strategic options expanded as you discovered more surface area on the world map. Today, a similar concept applies for AI-native applications gathering context. Applications that are able to capture a lot of useful context will end up being some of the most valuable companies.
- The Dark Ages of Gathering Context --
In 2026, we are currently navigating through the dark ages of gathering context. On one end, companies realize the paramount importance of gathering context; however, most AI-native SaaS products are particularly lousy at that.
Let's take notetakers as an example. In some ways, notetakers have access to the richest source of context within a company. Access to conversational data across the company gives notetakers the ability to take on some really fascinating new surface areas, ranging from custom automations to proactive recommendations. To do that, however, notetakers would have to gather a bit more context on the bigger picture surrounding the company's goals, or the user's goals.
For AI applications, gathering the right context is the most crucial first step. The real size of the opportunity that applications can build around ultimately depends on the accessible context window. In many ways, data and context will become the core drivers of enterprise value in the AI age. And that prompts an important question -- how should companies think about what context to gather, and how to go about it?
- The "What" and "How" of Gathering Context --
In my view, the right strategy to gather context is to think through the lens of "synergistic data assets". Based on the context that the application has today, what additional data could make the bigger picture more evident and specific? What additional data points can make the current accessible context way more useful?
The path to gathering that new data or context relies on serving customers with value that justifies the effort required to provide that context. And the easiest way to do that is to build the software that would use the new piece of context to deliver value. And then ask users to integrate or ingest the missing context, using the newly developed capabilities as the reward.
In many ways, AI applications are only as strong as the underlying access to context.
Published on September 24th, 2026.