Agentic AI: Context is Key for Business Goals

Agentic AI needs to understand context better to help businesses reach their goals. A CEO from the Swedish vendor, Orient, states that agents need to understand us better to fix problems caused by poor AI use. Many transactional benefits of automating tasks, like writing reports, are lost if you spend hours checking the output. This creates more work than it removes, leaving employees verifying, correcting, and contextualizing. Orient claims this is why organizations struggle to see a return on investment from AI. Research from Stanford University found that even AIs trained on trusted data sets invented citations. For example, some legal AI systems produced incorrect information over 17% of the time, and one hallucinated more than 34% of the time. Orient believes its agents can turn work into "structured meaning objects" that include claims, sources, and evidence. This allows agents to reason from context instead of guessing from files. Their website explains that organizations need better ways to know what's true and trusted, and what people understand. This approach helps agents act with context and evidence.