EmbeddingGemma 2: an open, lightweight multimodal embedding model
deepmind.google·6d ago
AI Engineering
MIT Technology Review·Aug 12·5 min read
Scaling AI agents with trustworthy data
TL;DR
Organizations are increasingly adopting agentic AI, but many struggle to achieve ROI due to inadequate data infrastructure. Data leaders, who provide broader access to enterprise data, are seeing greater success with AI agents.
✦ Why It Matters
Evaluate your data access policies today to ensure AI agents can operate effectively.
Key Takeaways
- Assess your data infrastructure for agentic AI readiness.
- Aim for over 70% data access to improve AI performance.
- Trust in AI decisions correlates with data availability.
- Focus on automating data management for better governance.
Full Summary
Agentic AI, which shifts from merely answering questions to taking actions, requires comprehensive access to both structured and unstructured data across an organization. A recent survey of 300 data and technology executives revealed that AI agents only access an average of 45% of company data, with data laggards accessing even less.
In contrast, data leaders ensure over 70% access, leading to higher trust in AI decisions—100% of data leaders trust their agents compared to only half of all respondents. Legacy data systems are a significant barrier, with two-thirds of data laggards citing limitations in scaling AI agents.
As organizations aim to implement agentic AI widely within two years, improving data access and governance is crucial for success.