---
type: digest
source_type: x
category: reference
domain: enterprise-adoption
status: captured
captured: 2026-06-01
published: 2026-06-01
confidence: medium
tags:
  - aaron-levie
  - enterprise-adoption
  - agentic-ai
  - context-engineering
  - knowledge-management
  - workflow-migration
  - access-control
source: https://x.com/levie/status/2061247380897579500
raw_source:
  - "[[00 Raw/X Posts/2026-06-01 Aaron Levie - Enterprise Agents Context Problem]]"
---

# Aaron Levie - Enterprise Agents Context Problem - Digest

## Core claim

Aaron Levie argues that the number-one problem for enterprise AI agents is context: knowledge-work agents need business context that is often fragmented across legacy systems, blocked by mismatched access controls, or never digitized because it lives as tribal knowledge in people's heads.

## Useful KB signal

- Agentic coding is context-rich compared with general knowledge work because codebases encode much of the task state and technical users can supply missing detail.
- Enterprise knowledge work is context-poor because decisions, processes, workflows, domain expertise, and document history are scattered or tacit.
- Access control is not only a security issue; it is a workflow fit issue. Agents need permission models that map to the real work being done.
- Modern, secure cloud knowledge environments become part of the agent deployment substrate.
- Applied AI companies, FDE teams, and system integrators can create leverage by capturing, structuring, permissioning, and delivering domain context to agents.

## Strategic interpretation

This strengthens the KB's view that enterprise agent adoption is fundamentally a context and workflow problem, not just a model-capability problem. The hard work is turning fragmented digital knowledge and undocumented institutional practice into governed operating memory that agents can use safely.

The most valuable enterprise-agent vendors may therefore be the ones that solve the context supply chain: where knowledge lives, who can access it, how tacit decisions become retrievable data, how workflows map to permissions, and how the agent receives exactly enough domain context to act.

## Caveats

- This is a high-signal operator field note from Levie's Box customer work, not a measured survey.
- The post is most useful as directional adoption signal and should be paired with concrete enterprise deployment evidence around knowledge migration, access governance, and workflow outcomes.

## Links

- Raw source: [[00 Raw/X Posts/2026-06-01 Aaron Levie - Enterprise Agents Context Problem]]
- Related: [[05 People/Aaron Levie]]
- Related: [[03 Research/Enterprise Adoption/Enterprise Adoption Overview]]
- Related: [[03 Research/Agentic AI/Agentic AI Overview]]
- Related: [[06 Theses/The Workflow Migration Thesis]]
- Related: [[06 Theses/The Control Surface Thesis]]
- Related: [[09 Implications/Treat Memory and Identity as Infrastructure]]
- Related: [[09 Implications/Pair Governed Runtimes With Workflow Owners]]
