NewsBite analysis · AI, Business, Cloud
VB
AI agent infrastructure: 3 enterprise fixes
During a recent event, enterprise leaders from Walmart and Zendesk highlighted challenges they faced integrating AI agents into existing infrastructure. They noted that traditional setups are ill-suited for AI-driven workflows, often leading to bottlenecks during scaling. Key strategies discussed included investing in robust evaluations, owning AI harnesses from day one, and ensuring model independence to maintain flexibility in approaches as AI technology evolves. These insights underscore a growing recognition of the need for governance and optimization in leveraging AI effectively.
MonitorConfidence: HighMid term horizonData governance risk: High
Published 17 Jul 2026, 19:25 UTC4 min read
Visit original article ↗Lead image — VB
Summary
TL;DR
- Walmart built a proprietary internal gateway to manage AI agents efficiently.
- Zendesk faced data bottlenecks due to a vast repository of customer interactions.
- Governance strategies were essential for managing overlapping AI projects.
- Investing in evaluations is critical to speeding up AI deployment.
- Flexibility in model choice enhances the effectiveness of AI solutions.
Interesting facts
- Walmart's internal agent harness became popular among employees, leading to citizen-developed solutions.
- Zendesk's discussions included managing billions of customer conversations efficiently.