By Ishya Anbuselvan
Gartner predicts that over 40% of agentic AI projects would be canceled by the end of 2027. The research firm attributed this to rising expenses, vague value propositions, and technology firms failing to manage risk adequately. [1]
These three factors are severe barriers to deploying AI agents effectively, and such factors make it challenging for government leaders to correcting analyze the real costs and complexities of rolling out scalable AI agent. When federal project or technology teams want to increase AI use in agencies, they often skip crucial steps which are needed to successfully deploy an agent.
The end result is that AI agents stay stuck in the pilot phase, leaders cannot scale their solutions, or they buy tools that are not even real AI agents.
According to Gartner, the AI hype has led to “agent washing” in the market. This is where AI contractors and technology providers rebrand existing products, such as AI assistants, robotic process automation (RPA) and chatbots, without substantial or real agentic capabilities. Gartner estimate that only about 130 of the thousands of AI agentic vendors out there are actually selling legitimate agents, which is only10% of the 1,414-vendor figure reported by the data tracking company Tracxn. [2]
The statistics highlight a challenge for federal technology leaders. In order for technology leaders and federal customers to get true value and access to the latest technology to achieve their mission, they must understand the difference between true AI agents and “agent washing.” There’s a key difference between superficial chatbots and coworking platforms that possess agentic capabilities. Marketing a simple chatbot with the “agentic” label does not give it the capacity to reason critically, pivot in the face of challenges, work securely within federal customers’ preexisting structures, or provide mission impact.
Why Do AI Agents Matter?
AI agents are the future of AI and more importantly where organizations are able to realize real value from AI. Dr. Andrej Karpathy, an AI researcher and founding member of OpenAI, recently said that we are in the decade of agentic AI. AI agents have the ability to act autonomously in the background to autonomously achieve small tasks. Simply put, think of AI agents as smart interns, that are given direction and then work autonomously to achieve administrative tasks, like reviewing invoices, doing market research, assisting with human resources onboarding, and analyzing contract clauses. The end result is that a federal employee’s productivity is increased, so he or she can work on more important tasks. In a time when so many agencies face budget cuts and have had staffing cuts, AI agents are a needed solution.
Avenir Technology: A Firm Developing Real Agents
The AI landscape is changing rapidly, and more users and agencies are using agentic developer tools and web assistants like ChatGPT, Claude’s Opus, and Gemini Pro. But these tools are fundamentally chatbots requiring a lot of manual prompting to execute complex tasks. True agentic AI does not require a user to guide all of its actions because agentic AI means there is more autonomy. Think of chatbots as being reactive and agentic AI is more proactive in actually doing and managing tasks and not requiring prompting.
Today, Avenir Technology announces the launch of Avenir Orchestra™, a platform for the federal government that offers real agents that function as a highly adaptable coworker. While designing Orchestra, Avenir’s team wanted to develop a solution beyond the hype and instead implement a platform and features that will benefit federal leadership and employees and imbed and assist with their workflows. Best of all, it can be customized for federal agencies to assist with different workflows and needs.
Avenir Technology’s CEO, Ashley Nicholson, said of Avenir Orchestra™’s launch that “everyone is talking about AI agents. It’s become a buzzword..but it’s shocking how most of these companies are not deploying real agents in real, mission-focused environments. Also, they are not battle tested or developed with federal agencies in mind. Our team built Avenir Orchestra™ and have tested it and deployed it for commercial customers. We have lessons learned and have refined the platform for federal customers. Also, our team understands federal AI compliance and security and also best practices, like human-in-the-loop approaches to implementing agents responsibly, so we are able to manage AI agent governance and compliance effectively. As the Gartner research shows, there are very few in the market who can say this.”
AI Agents in a Government Setting
When emerging technology are rolled out for public sector clients, new concerns that go beyond a tool’s agentic capabilities arise, due to government’s need to serve citizens equitably and the higher security and compliance required. Government agencies want to harness the power of AI agents and increase productivity, but are concerned about security, safety, and compliance with many laws and regulations. Government customers ask themselves: where does this system actually run? How do I ensure my data is secure? How can I ensure that AI has guardrails?
Unlike other providers, Avenir’s team has built Avenir Orchestra™ with the federal government in mind. Our team has over 100 years’ collective experience and we have drawn on this expertise when building the platform integrations and security features. Our team is aware of government requirements, FedRamp requirements, and federal cybersecurity regulations. Also, our team has proven approaches and tools to ensure that AI agents are deployed responsibly according to Executive Order 14409 “Promoting Advanced Artificial Intelligence & Innovation Security,” Executive Order 14179, Removing Barriers to American Leadership in Artificial Intelligence”, and the OMB Memorandum M-25-21 “Accelerating Federal Use of AI through Innovation Governance and Public Trust, and the NIST AI Risk Management Framework.
Avenir Orchestra™’s Security Features
Avenir Orchestra™ platform helps government agencies and leaders ensure that AI agents are deployed responsibly. Our team has built security features into the Avenir Orchestra™ platform. The platform can be securely hosted within FedRAMP cloud environments in AWS or Azure, and agencies are free to use whichever LLM or foundational models which fit their needs and according to agency policies.
Orchestra is also built on a customizable, industry standard, LangChain architecture, prioritizing clear and hardcoded security over the lax security protocols typical of low- or no-code software like n8n or OpenClaw.
Our team has operationalized compliance with federal regulations, incluiding Executive Order 14409 “Promoting Advanced Artificial Intelligence & Innovation Security,” Executive Order 14179, Removing Barriers to American Leadership in Artificial Intelligence”, and the OMB Memorandum M-25-21 “Accelerating Federal Use of AI through Innovation Governance and Public Trust, and the NIST AI Risk Management Framework. We have achieved this by building immutable audit trails, ensuring malicious actors cannot tamper with system logs, and human-in-the-loop checks into the platform. Additionally, we assist clients with building AI governance plans and policies. This ensures that agents are deployed responsibly, can be monitored, and that there are guardrails in place.
Moving Past Experimentation: The Importance of Experience with Real World Deployments
Federal leaders often see impressive tech demos on how an AI agent operates under perfect conditions. Deployment in a client’s real day-to-day environment, however, looks very different and the technology fails to deliver. This is always a risk for federal customers. According to Gartner, unlike many technology companies in the space, Avenir Technology is the 10% of companies that have real world experience actually deploying AI agents and scaling them up. Avenir has had extensive commercial deployments across several mission-critical workflows and such experiences has allowed our team to refine and improve our platform. Avenir’s team, as a result of learning from real-world deployments, has optimized Avenir Orchestra™ for the federal government and has approaches to minimize hallucinations and model drift.
Avenir Orchestra™’s Impact and Value
In Avenir’s commercial deployments, there has been a tremendous impact and return on investment. Employees at a company that deployed AI agents using the Avenir Orchestra™ platform were able to reduce administrative tasks by 25%. Also, the leadership at that organization told us that this has increased employee morale because they are able to work on higher level tasks and focus on the mission. In another deployment, our team was able to reduce a HR team’s administrative burden by 20%, which allowed the HR staff to focus on professional development, assisting leadership, compliance, and other higher-level tasks.
To learn more about Avenir Orchestra™, visit our website visit our webpage at www.avenirtechnology.ai or contact us at info@avenirtechnology.ai.
Avenir Technology is a dynamic technology company that helps private and public sector clients make better and data driven decision to support their missions. With an extensive background in technology, our team offers data analytics services, AI services, digital services, and emerging technology services. We are an SBA woman-owned small business (WOSB) and 8(a) company and our team has over 100 years’ experience in support government technology. Avenir Orchestra™ is available for organizations and agencies interested in exploring customized and secure AI agent solutions.
[1] Gartner Predicts Over 40% of Agentic AI Projects Will Be Canceled by End of 2027 [2] Gartner Predicts Over 40% of Agentic AI Projects Will Be Canceled by End of 2027