Public sector organizations are turning to purpose-built small language models (SLMs) as a practical solution for operationalizing AI. These models offer the security and control that government agencies require, while avoiding the constraints of larger models.

The Capgemini study highlights the 79 percent concern among public sector executives about AI’s data security. Government agencies must ensure their data stays under their control, with minimal disruptions to operations—an impossible task with large language models (LLMs).

SLMs are specialized to meet departmental needs, using billions rather than hundreds of billions of parameters, making them computationally less demanding and more secure. They can be housed locally, ensuring continuous operation without relying on cloud connectivity.

These models offer superior search capabilities, revolutionizing how the government manages unstructured data and drafts complex texts legally compliantly. AI’s potential extends beyond chatbots to interpret data and support executive decision-making.