Why context graphs might be the next big thing

A 'context graph' explains WHY something happened.

Existing systems often document WHAT happened - a price changed, a box moved, a team relocated. They miss the decisions and tradeoffs which lead to that event happening.

Although an accurate reflection of what happened. It gives a massive data hole for Agentic AI, to understand how the organisation got to where it is.

Exceptions are common and are often a point of practice, not policy. We always give [x] segment an extra 30 days payment terms over the winter period.

They may a precedent from past decisions, where the same is replicated whilst the situation has moved on.

Or a product of the conflict, or compromise between two or more teams in our complex, cross-functional environments.

Without context, it is hard for Agentic AI to learn (and replace) the humans. It is missing the data it needs to make the calls. Data that doesn't exist anywhere except in your team’s head.

Which is where the problem exists. Past decisions are unlikely to be tracked back, but future systems can make more sense of the data that they have and the richness of conversations that lead to a choice.

For Agentic AI to replace the humans (HOW), it needs to understand the WHY, not just the WHAT.

Source
Article: https://foundationcapital.com/context-graphs-ais-trillion-dollar-opportunity/?_bhlid=78d20398e79cce9695aa51615b1f53aec09408f6

BESCI AI OPINION

This is the messy truth behind AI - that so much of why an organisation does things - is simply not captured.

Will it be - yes, because it will need to be if the aim to replace the humans with Agentic AI and a much lower cost drives it. The question will be how.

With access to transcripts, emails, Slack messages and teams chats, an agent might be able to work out the context behind the decisions being made. A deep data pool exists, if you know where to look.

It only takes one person to offer the context and for it to be documented. Which reminds me of this example in the medical billing industry, where a billing analyst was hired by two entrepreneurs - typically stuck in a windowless, basement office, they saw the value in what she knew, and boy was it good. (https://www.pushkin.fm/podcasts/against-the-rules/six-levels-down)