Blog
Over the years working in government IT, I’ve often had to temper the excitement that accompanies emerging technologies with the realities of deploying them across sensitive, complex environments.
This tension has only intensified over the past few years as AI and LLMs exploded on the scene. From the outset, AI has offered a tantalizing opportunity to revolutionize government operations to be more efficient and agile, improve citizen services, and ultimately contribute to America’s competitiveness.
In my entire career, I’ve never witnessed anything so transformational and I’m far from alone. Across the federal landscape, AI is everywhere. It dominates public sector media, networking events, technical meetups, customer conversations, and many of the most tantalizing solicitations. Its momentum shows no signs of slowing.
The Federal Government’s AI Manifest Destiny
The urgency for federal entities to adopt AI technologies has never been more pronounced. It’s not simply about transformation and modernization, it’s a global economic and geopolitical competition, underscored by OMB Mandates M-25-21 and M-25-22.
Collectively, the OMB memos signal a move away from a risk-averse stance toward a more streamlined approach that embraces innovation, promotes fail fast commercial mentalities, and is designed to drive the United States’ dominance in AI development.
Yet the path forward is not without its challenges. Concerns around data protections, security, and lack of standardization are real and valid. That’s why the OMB’s emphasis on responsible AI adoption is critical. By calling for governance frameworks, guardrails, and the designation of Chief AI Officers across federal agencies, the directives establish a clear framework for thoughtful implementation and operations.
The Agentic AI Wild West
Perhaps one of the most interesting takeaways of this new AI era is that much of the innovations are happening in a transparent manner with open-source collaboration. This has been a lynchpin for the AI community to identify and address foundational standardization. Remember the Protocol Wars? Those debates and battles over communications protocols played out over decades. Now, standards are being proposed, debated and adopted in the open over a matter of months, weeks and even days.
A prime example: In November 2024, Anthropic introduced the Model Context Protocol (MCP) which was conceptualized as a standardized communication layer for LLMs. Think of it as the USB-C of AI: a universal way to plug applications into LLMs, enabling consistent context sharing and connectivity across tools and data sources.
The adoption was swift. By early 2025, more than 10,000 MCP servers were operational. GitHub activity surged. Tech giants like AWS, Google, OpenAI, and Microsoft quickly integrated MCP into their ecosystems and client-side tools like Claude Desktop, Claude Code, and Cursor all rolled out MCP integrations.
Having a standardized protocol to connect LLMs into valuable ecosystems quickly enabled developers to build AI Agents, which essentially unleashes the power of GenAI into a wealth of new use cases. They can operate collaboratively with users or autonomously on their behalf. Their functions range from answering straightforward questions to executing complex, multi-step tasks. Unlike traditional personal assistants, agents can be customized with domain-specific expertise to accomplish mission tasks 10-100x faster.
The result has been an explosion of AI Agents into the ecosystem and while they represent a significant opportunity for the Public Sector that is looking to create efficiencies, reduce costs and streamline operations, they also create significant new security risks.
The Sherrif Arrives
The security concerns of plugging these agents into sensitive government systems poses a major roadblock for Public Sector agencies to adopt AI. The potential for data leakage, system compromise, and misaligned behavior is significant.
Fortunately, our partner Lasso Security stepped into the void and in mid-April, unveiled the Lasso MCP Gateway, an open-source proxy and orchestrator for all MCP interactions, embedding critical security, governance, and monitoring capabilities while maintaining full compatibility with existing agent implementations.
It can’t be understated how critical Lasso’s MCP Gateway is for Public Sector agencies. It epitomizes the kind of innovation that aligns with the government’s dual mandate of progress and protection.
With Lasso’s MCP Gateway, agencies can adopt MCP connected AI agents with confidence, knowing they have real-time visibility, AI-powered LLM firewall protection, and a NIST AI RMF aligned architecture built in.
Swish Secures IT
Swish is proud to be the exclusive provider of Lasso Security’s AI solutions to the federal market. Our unique expertise in generative AI solutions, cybersecurity, and observability enables us to guide agencies through the complexities of GenAI platform deployments with security and performance at the forefront.
For government agencies exploring generative AI, understanding MCP and adopting solutions like the MCP Gateway isn’t optional, it’s essential. This approach not only aligns with OMB mandates and NIST AI guidance but also lays the foundation for safe, scalable, and mission-aligned AI integration.
As we stand on the cusp of a new era in government technology, the dialogue around GenAI’s potential versus its perils becomes increasingly pertinent. As facilitators of this journey, we at Swish are committed to ensuring that government entities are equipped with the knowledge and tools necessary to navigate this complex landscape.
The road ahead is both exciting and challenging. Together, we have the opportunity to redefine what is possible, setting a new standard for innovation, security, and public service in the age of AI.
If you’re navigating the next wave of AI adoption in the public sector, let’s talk!
