Thoughts on AI in Enterprise Workflows
2026 begins with AI deeply embedded in enterprise operations. Here's what we're seeing and where it's heading.
Current State
- LLMs augmenting knowledge work
- Computer vision in quality inspection
- Predictive models for demand planning
- Conversational interfaces for systems
Effective Patterns
- Human-in-the-loop: AI proposes, humans approve
- Narrow focus: Specific tasks, not general replacement
- Measurable outcomes: Clear success metrics
- Fallback paths: Graceful degradation when AI fails
What's Not Working
- Overpromising autonomous capabilities
- Ignoring data quality requirements
- Underinvesting in change management
AI is a tool. Success comes from thoughtful application to specific problems.