Semantic processing
LLMs interpret evidence through typed schemas and plain-language definitions. The core handles structure, identity, and authority while the schema supplies business meaning.
THE VISION
An enterprise that
learns without losing itself.
As applications and AI models change, the enterprise should retain what it knows, what that knowledge means, and the basis on which it acts.
Watch the foundation take shape
An illustrated explanation of why trusted context matters for AI, using the Fleks sales scenario.
Watch on YouTubeDownload the video (MP4)
Read the transcript
Introduce Fleks Fleks extracts data and context from everyday interactions, structured and governed by your data models. Creating the trusted foundation to unlock the power of AI.
The foundation AI needs AI promises better answers and increasingly capable automation. But unreliable context can turn a wrong assumption into a confidently wrong answer, or an action with unintended consequences. Realizing AI's potential starts with a foundation you can trust.
Make it concrete Maya, a busy sales rep, has a discovery call with a prospect. Fleks identifies the people, organization, and tasks they discussed, connecting them into a FleksScene.
Establish trust Maya checks the source evidence, makes corrections, and accepts the facts.
Adapt with the business Maya spots a gap. Fleks captured a task to define success criteria, but she needs to track the whole proof of concept project, including its sponsor, organization, and related tasks. She tells Fleks what's missing. Fleks proposes a new object in her data model for her ops team to review and approve.
Create trusted business records With the new object approved, Fleks adds the proof of concept project to Maya's FleksScene, connecting it to Northstar, Alex, and the success-criteria task. Maya reviews and accepts the new facts and relationships, creating a trusted business record with its supporting context.
Bridge to systems of record Fleks provides trusted business records along with their supporting context to the applications and systems that power your business.
The opportunity Your business. Your data models. Fleks provides the foundation for turning trusted context into governed action.
The next source of enduring value
AI is changing the effort required to create and adapt software. The lasting advantage may increasingly come from the knowledge an enterprise accumulates—and its ability to use that knowledge responsibly.
Today, context is scattered across conversations, documents, applications, and the working memories of people and agents. Each system can reconstruct a partial picture. The enterprise needs a basis for understanding that survives beyond any one of them.
From the White Paper: Sections 1–2: changing economics and trusted context.
Give enterprise understanding an architectural home
Fleks proposes an independent trusted knowledge plane. Applications and agents can contribute evidence, propose changes, and work from accepted context. Knowledge, definitions, and decision history remain part of the enterprise’s foundation as those participants change.
Applications should build on enterprise understanding. Enterprise understanding should outlast its applications.
This is a shared logical layer. It can coexist with distributed deployment, organizational boundaries, and distinct decision authorities. Its purpose is to make accepted context coherent and accountable across work.
Three capabilities create the possibility
Dynamic schema
The enterprise can adopt new concepts as its work changes, while preserving the meaning and lineage of previously accepted records.
Learning from history
Accumulated evidence and reviewed outcomes can inform better proposals and policy evolution. Systematic learning remains a capability to develop and evaluate.
From the White Paper: Section 7.2: early progress toward a new era.
A broader set of possibilities
People could spend less time reconstructing context. Applications could adapt around an organization’s own language. Agents could work from explicit, reviewable foundations for action. Enterprise knowledge could compound as software changes around it.
These are the ambitions of The Fleks Project. Its working POC gives us mechanisms to examine; the larger benefits need implementation, independent evaluation, and experience in real organizations.
From the White Paper: Section 8: beyond SaaS.