AI in Acquisition: Transforming Federal Procurement with Human-Centered Innovation

How artificial intelligence is reshaping government contracting while keeping humans at the center of decision-making

Read the full report on Leveraging AI for Transformative Federal Acquisition

Artificial intelligence is rapidly emerging as a game-changer for federal acquisition, promising to tackle the persistent challenges of slow procurement cycles, overworked contracting officers, and the complexity of navigating regulatory environments. In our latest webcast, I sat down with Christopher Barlow from MITRE and Wilson Miles from the National Defense Industrial Association’s Emerging Technologies Institute (ETI) to unpack how AI is already making a difference-and what it will take to unlock its full potential in government contracting.

Chris brings deep expertise from MITRE’s Acquisition Innovation Center, where he’s focused on AI strategies and practical tools for transforming procurement. Wilson’s research at ETI zeroes in on the intersection of emerging tech, supply chains, and acquisition policy, making him a leading voice on modernization challenges and workforce issues.

Together, we explore the urgent need for speed in federal acquisition, spotlighting how AI can accelerate market research, automate contract drafting, and streamline compliance-while also addressing the critical barriers of cultural resistance, fragmented data, and the ever-present need for experienced human decision-makers. We also dive into the skills gap and workforce planning, discussing how agencies can upskill teams and ensure knowledge transfer as technology evolves.

---

Conversation Deep Dive

The integration of artificial intelligence into federal acquisition processes represents a transformative opportunity to address longstanding challenges of bureaucratic inefficiency, overworked contracting officers, and extended procurement timelines. This article explores how AI can augment human capabilities throughout the acquisition lifecycle, the real-world applications already showing promise, and the critical balance between technological advancement and human expertise.

The Acquisition Challenge: A System Under Strain

The federal acquisition system, particularly within the Department of Defense (DoD), faces mounting pressure to modernize in the face of rapid technological change and evolving threats. Traditional procurement methods-characterized by manual reviews, duplicative paperwork, multiple approval layers, and inconsistent data management-have resulted in extended acquisition timelines that undermine mission readiness.

According to the Government Accountability Office, the DoD takes an average of 309 days to award complex service contracts due to administrative bottlenecks and fragmented information systems. In some cases, as Chris Barlow noted during our webcast, the lead time for major system acquisitions can stretch to nearly two years.

"When mapping out those processes, we are seeing extended lead times," Chris explained. "When you talk to the program managers and they actually see what those lead times look like... some of them were very shocked."

This problem is compounded by workforce challenges. The DoD acquisition workforce, consisting of 157,594 members (both civilian and military personnel), is frequently described as overworked. Wilson Miles, emphasized this point:

"Contracting officers are incredibly overworked and understaffed. People often complain about how long the acquisition process takes, particularly for the Department of Defense."

The consequences extend beyond just bureaucratic frustration. Extended procurement timelines create strategic disadvantages compared to international competitors. As Barlow noted, "If you look at our adversaries and some other organizations internationally, they're able to acquire warfighting capabilities at a much quicker pace than we are."

The strain on the system also creates risk for vendors and contractors. Long timelines between contract pursuit and award create uncertainty for businesses-especially smaller, innovative companies-about whether they have the financial runway to sustain themselves through the process.

The AI Opportunity: Beyond Automation to Augmentation

Artificial intelligence offers a promising set of tools to address these challenges by shifting mundane work away from acquisition professionals to IT systems. For the purposes of this discussion, AI encompasses several technologies, including machine learning, generative AI, retrieval augmented generation, multi-modal systems, and robotic process automation.

The potential benefits are substantial. A 2023 MIT study found that using generative AI tools like ChatGPT substantially raised productivity: the average time to complete controlled writing tasks decreased by 40% while output quality rose by 18%. Applied to acquisition, AI can help deliver capabilities faster by streamlining processes and enhancing outcomes.

"This is really about delivering capabilities faster," Wilson emphasized. "One of the ways we're thinking about that problem is through speeding up that process using AI tools."

However, both Chris and Wilson stressed that the goal isn't merely speed, but enhanced quality. "Aside from speed to delivery, we're also looking at enhancing outcomes," Chris noted. "Leveraging AI from a background of being able to collect all of your organization's data, see what you've done in the past, analyze your previous outcomes-you should, in theory, be able to apply AI to enhance your outcome."

AI Across the Acquisition Lifecycle: Practical Applications

The acquisition lifecycle consists of several phases, each with opportunities for AI integration:

1. Needs Identification and Requirements Definition
   AI can help articulate program needs by analyzing previous contracts and identifying common patterns and language. It can map stakeholder interests and provide technically feasible starting points for requirements.

2. Market Research and Analysis
   This is often a lengthy step that AI could make more insightful by analyzing vast amounts of data to identify potential vendors, assess market trends, and predict future needs. Natural language processing can quickly extract relevant information from industry reports and databases.

"You know, we hear a lot that contracting officers are incredibly overworked," Miles noted. "One of the ways that we're thinking about that problem is through speeding up that process using AI tools, whether helping the Department of Defense to do a better job at conducting market research or writing contracts."

3. Acquisition Strategy Development
   AI can analyze data from past program strategies and outcomes to help determine the best acquisition approach. It can also educate program managers on why a particular strategy was chosen and what's required to execute it.

4. Solicitation Creation
   AI tools can automate the development of requests for proposals (RFPs) into standardized formats, resulting in quicker solicitation times. They can develop evaluation criteria based on large datasets of previous solicitations and responses.

5. Evaluation and Source Selection
   AI can assist in evaluating proposals against predefined criteria, analyzing supplier capabilities and past performance data. It can predict the likelihood of a vendor's success based on historical data, helping to provide more informed decisions.

6. Contract Award and Negotiation
   AI tools can suggest optimal contract terms and conditions by analyzing similar contracts and outcomes. They can detect anomalies and potential fraud in contract awards by analyzing patterns and flagging suspicious activities.

7. Contract Management
   AI can continuously monitor contract performance using data analytics, alerting managers to potential issues before they become significant problems. It can generate reports on performance, compliance, and financials, reducing administrative burden.

8. Contract Closeout and Evaluation
   AI can streamline the review of contract documents to assess whether all obligations have been met and identify outstanding issues. It can capture lessons learned and best practices from closed contracts, providing valuable insights for future acquisition strategies.

Early Success Stories: Pilots Showing Promise

While widespread adoption is still pending, early movers in federal acquisition have begun to see tangible benefits from initial applications. The Defense Logistics Agency has employed AI to optimize inventory management through supply chain forecasting and demand planning. The Air Force has explored AI tools for personnel and resource management, including the recent launch of NIPRGPT, designed to assist users with correspondence, background papers, and code.

In late 2024, the Army announced a pilot program experimenting with a generative AI tool to assist with multiple acquisition activities. The Army AI Integration Center also developed CamoGPT, which is built to optimize equipment maintenance, logistics, and supply chain management.

Miles shared an example from his LinkedIn network: "Someone posted about NIPRGPT, which is an AFRL (Air Force Research Laboratory) tool, and they were saying that it helped them put together a list of questions to present to vendors who are making proposals on a topic. It's just a really small win, but these informal successes are happening."

Human in the Loop: The Critical Balance

Perhaps the most crucial aspect of AI integration into acquisition is maintaining the right balance between technological capability and human judgment. Both experts emphasized repeatedly that AI should augment human capabilities, not replace human decision-makers.

"The human focusing on those outcomes gets you to the point where now all of your programs should have a better understanding of what the true problem set is," Barlow explained. "Staying in that strategic mindset should align you closer with that objective of the entire organization, where sometimes if you don't have the time to do that stuff, you're just going for what can I get based on the time that I have."

This human-in-the-loop approach is not just a philosophical preference-it's a practical necessity. As Wilson Miles noted, "In the DoD context, the level of review doesn't go away in acquisition just because there's AI. There's never not going to be a human in the loop at the end of the day."

The legal and regulatory framework reinforces this requirement. Certain functions are inherently governmental, which statutes and regulations define as tasks that must be performed by government officials. These include functions requiring discretion over governance areas such as policy decision-making, performance accountability, and execution of monetary transactions.

Chris Barlow illustrated this with a powerful example:

"If a contract officer leverages AI to build a contract, the contract becomes awarded, and upon award we realized that there were some data or IP clauses that were not included in that contract, and now the vendor owns government data-there's a huge risk to security within that system. Who's going to get in trouble? It's not going to be the AI."

The Skills Gap Paradox

One of the most thought-provoking discussions during the webcast centered on what might be called the "skills gap paradox." As AI takes over more routine tasks, there's a risk that junior professionals won't develop the foundational knowledge needed to eventually become experts.

"While it is entirely possible that some people's jobs may be at risk because of increased efficiency with AI, that's not the trend that we're seeing," Barlow noted. "What we are seeing is that you actually need more people because the workload is already overbearing."

However, he identified a significant risk in succession planning: "If we could supplement all of the junior-level work because we have the experts to refine all of that, then we don't need junior-level engineers. That's simply not true. You have a huge risk of succession planning if you are not bringing in junior people and getting them up to speed and training them and creating the opportunity for them to become those subject matter experts."

This highlights the need for a balanced approach to AI integration-one that automates routine tasks while still providing opportunities for professional development and knowledge transfer between generations.

Challenges to Adoption: Culture, Policy, and Technology

Despite the promising applications, several significant barriers stand in the way of widespread AI adoption in acquisition:

Cultural Challenges

The DoD's strong warfighter culture can make it difficult to justify spending on "back office" functions rather than weapons platforms. Additionally, there's often resistance from contracting officers who lack the time to learn new tools.

"Contracting officers don't have the time to learn a new tool," Miles explained. "It has to be extremely simple so that they don't have to take a class on it."

Fear of job loss and risk-aversion also contribute to cultural resistance. Approximately 50% of the DoD civilian acquisition workforce consists of individuals aged 40 and above, creating varying levels of comfort with new technologies.

Policy Challenges

While policy itself isn't necessarily a barrier-there's nothing in the Federal Acquisition Regulation (FAR) that prohibits using AI tools-there are important considerations around inherently governmental functions, classification of information, and intellectual property rights.

Current DoD policy restricts the use of national security information and controlled unclassified information (CUI) in publicly accessible AI tools. This necessitates the development of DoD-specific solutions that meet security requirements.

Technical Challenges

Data quality and availability represent significant technical hurdles. The DoD has struggled to collect and retain data about acquisition processes and program execution, and to share data across government and the private sector.

"DOD is both swimming in data as well as doesn't know what to do with that data, and they're also not good at collecting data," Miles noted. "There are these two sides of a coin."

The Authority to Operate (ATO) process, which determines when new software can be installed and used on government systems, is often the longest step in deploying software. This process is particularly challenging for small and non-traditional businesses developing innovative AI solutions.

The Path Forward: Recommendations for Implementation

Based on the research and expert insights, several key recommendations emerge for successfully integrating AI into acquisition processes:

Understand Your Workflow

Organizations should identify painful parts of their process where hours are spent on mundane and repetitive tasks, map these workflows, and challenge assumptions about why particular steps exist.

"Pick a painful part of your process where hours are spent on this really mundane and repetitive task, map it, understand in your workflow why that particular step exists, and challenge it," Miles advised.

Pilot Programs with Clear Metrics

Agencies should establish pilot programs using commercial AI tools for select applications in the contracting lifecycle, with strict criteria to evaluate success and plans to scale successful tools.

"We're fully for more pilot programs," Miles stated. "The best way is for the Hill to push the department to use more commercially available AI tools. There needs to be strict criteria for evaluating success."

Focus on Process First, Then Technology

Organizations should streamline processes manually before adding AI enhancements. As Barlow emphasized, "These tools should be enhancements to systems, not problem solving inherently. If that process is not efficient to start with, throwing a tool at it is very likely to not be as successful as you're hoping it to be."

Build AI Literacy Through Training

Comprehensive training programs should be developed to ensure AI literacy across the workforce, with varying levels based on roles and responsibilities. This includes understanding both the capabilities and limitations of AI tools.

Invest in Data Infrastructure

Significant investments in infrastructure and robust data governance policies are necessary for AI adoption to succeed. This includes establishing standards for data collection, usage, and sharing.

Plan for Success, Not Just Failure

Organizations need to think beyond the pilot phase and plan for what happens if AI tools exceed expectations. As Barlow succinctly put it: "Plan for failure, but prepare for success."

Conclusion: A Human-Centered Technological Revolution

The integration of AI into federal acquisition represents not just a technological shift but a fundamental reimagining of how government procures goods and services. By automating routine tasks, enhancing decision-making, and improving outcomes, AI can help address the chronic challenges of overworked staff and extended timelines.

However, the most successful implementations will be those that recognize AI as a tool to augment human capabilities rather than replace human judgment. As our experts emphasized throughout the webcast, the goal is to free acquisition professionals to focus on strategic thinking and complex decision-making while leveraging AI to handle the routine and repetitive aspects of the process.

The path forward requires a balanced approach-one that embraces technological innovation while preserving the essential human elements of the acquisition process. With thoughtful implementation, clear metrics, and a focus on building both technical infrastructure and human capability, AI can help transform federal acquisition into a more efficient, effective, and responsive system.

As we navigate this transformation, the guiding principle should be that AI exists to serve human needs and objectives-not the other way around. By keeping humans in the loop and focusing on outcomes rather than just processes, we can harness the full potential of AI to deliver better value for taxpayers and enhanced capabilities for those who serve our nation.