Preparing for AI in Acquisition
BY AL MUÑOZ, FELLOW
I am old enough to remember when the first laptop computers were delivered to the office. They were big and expensive and had no battery life at all.
After a brief “ooh and aah” moment with the office’s leadership, we promptly put them in the office safe. The laptops stayed there almost completely unused for a few years, next to the office’s single mobile phone. No one knew what to do with them. We could all see that they gave us capabilities that we hadn’t previously had, but we didn’t have a plan to assess those capabilities or to integrate them into our work. So, they sat, kept in reserve, in case something came up.
Not long after, a massive storm hit that knocked out the power for several hours and the local roads for days. The technology that would have kept us in operation was safely tucked away where no one could get to it until it was no longer needed again. Although those laptops had missed their moment, we didn’t grasp at the time how integral to our daily lives their progeny would eventually become.
Since then, there have been other waves of capabilities that have made their way to us that changed the way we work. Most recently, the talk of the town is the topic of artificial intelligence (AI). Unlike those laptops, AI will not be sitting in the safe until we figure out what to do with it. It is already here, in use, and agencies and industry have already turned their attention to it. It is time for those of us in the contracting profession to do so as well.
Automation and the Coming of AI
AI has the potential to replace large swaths of repetitive work with automation. (1) At first glance this appears to be the next chapter in a lengthy novel of attempts to automate what has typically been an automation-resistant field. Although the names of the systems change, we do in contracting essentially the same thing we’ve always done: we identify a need, find it, and buy it, which hasn’t changed since 1776.
Attempts to automate the process have mostly focused on capturing the process, rather than improving it. We need more contracting professionals than ever to help meet the need for speed. Even with current technological advancements that we as a society have experienced during the past two decades, procurement lead times have stubbornly stayed the same. So, what gives?
Where AI differs from automation is in how it works. (2) While process automation generally means taking existing rules-based processes and programming them into software that then follows the rules it is given, AI is designed to simulate human thinking.
The best automation will do what it is told until it is told to do something different, whereas AI systems can learn and adapt. Automation uses software, scripts, or robotics to perform predefined tasks; AI can make predictions, recognize patterns, and improve its performance over time.
With automation we can bake in gaps that need workarounds and other patches to remain useful. AI can evolve while it’s functioning, handling complex tasks that may involve ambiguity and uncertainty. (3)
Government procurement is highly repetitive and rules-based, things that AI already excels at improving. For many years we have envisioned a world where the mundane and time-consuming portion of the job is reduced or eliminated, and contracting professionals can focus on using their business judgment and safeguarding the country’s many interests.
As Figure 1 shows, some of the things we do are well within the capability of AI. With AI, the potential for moving the needle on contracting is again on the rise and has, in fact, already been making inroads into procurement processes both inside and outside government. (4)
Preparing the Profession for AI
Even though AI can do many things better and faster, it can’t do everything and probably never will. Tasks in the future will be enhanced by AI more than they will be replaced by AI.
Contracting professionals will remain in the equation, but their role will shift more dramatically than it ever has. AI can learn, it can reason, it can problem solve, and it can make decisions, all of which have always been part and parcel of the contracting profession, regardless of what was being purchased. Contracting professionals now need to focus their attention on developing skills that bring value to the acquisition process that AI cannot, including:
• Develop strong critical thinking and problem-solving skills. AI can handle routine tasks, but humans are still needed to tackle complex and unique challenges.
• Be creative. AI can assist in generating ideas, but human creativity is essential for delivering insights and making connections. (5)
• Hone interpersonal skills, emotional intelligence, and empathy. The professional categories that are furthest from AI integration include managing others, providing expertise, and interfacing with stakeholders. (6)
• Develop leadership skills. As AI becomes more widespread, leaders and teams that guide AI strategies will be in demand.
• Hone your knowledge. Deep subject matter expertise will set you apart from AI systems that can’t develop experience and expertise. AI used by subject matter experts will always be better than AI alone. (7)
Where AI will be the most immediately useful will be in tackling the distractions in the process, not in replacing business judgment and expertise. AI can spot patterns in the market and quickly inform buyers where bargaining might be had, but it will not do that bargaining for you.
AI can automate the process of justifying contracting decisions, but it won’t be able to spot whether those justifications are sound or simply comply with the rules. The AI-enhanced contracting world will still depend on the well-prepared professional, using their hard-gained expertise to balance the many contracting challenges effectively.
Preparing the Acquisition Office for AI
In addition to developing yourself, agencies and their stakeholders will need our input and participation to plan and execute successful AI integration. Like past improvements, it will not be an immediate change to the profession, but rather a slow period of incubation, which we appear to be in now. We will see rapid changes as we automate processes, then integrate AI recommendations into our work. Although each organization will move at a different pace, all organizations can expect to undertake certain tasks to prepare to incorporate AI. (8) These include:
• Assess the current state. Organizations will need practitioners to evaluate current workflows, systems, and capabilities to under-stand where AI can add value.
• Set clear objectives. We’ll need to help define clear goals and objectives for integrating AI in the acquisition shop and its processes. We’ll need to identify the operational and oversight problems or tasks we want AI to address and establish the outcomes we expect.
• Develop an AI strategy. Every organization will need to create an AI strategy that aligns with its business goals. This will be a good opportunity to assess the organization’s structure and assess if and where AI enhances business value for clients. (9)
• Assess and develop an AI skilled workforce and culture. We’ll need to invest the time and funds into training the acquisition workforce. We’ll need to ensure that we and our employees have the necessary AI literacy and skills to work effectively with AI systems.
In addition, there will be many steps to integrating AI into the craft and many factors to consider along the way – factors such as identifying data sources, working through ethical and security issues, and eliminating historical biases.
In an AI-enhanced world, the ever-watchful eyes of the acquisition workforce will need to adapt to the speed and power that AI will introduce into processes and outcomes. We’ll need to understand the many complexities of AI and be able to convey them effectively to our stakeholders and partners.
The Future is Here...Almost
To be fair, the current state of AI adoption is mixed and so far, the biggest wins have not moved the needle all that much. Although we have had some wins in robotic process automation and machine learning, we are not quite ready for the full implementation phase of AI into acquisition. There is still time to prepare yourself and your organization for the next set of advancements and to smooth what looks to improve and enhance a profession we all care strongly about.
Perhaps the first thing we need to recognize and respond to is that AI isn’t waiting for a rainy day. It is only a matter of time before we will need the skills and savvy required to make AI an integral partner in the profession.
We’ve all experienced the many changes in the craft coming faster and faster and we’ve always been up to the task of changing with it using our adaptability and expertise. AI will never replace us, but the future acquisition workforce with AI will. It is time to prepare to be a part of that future workforce and give it the value that only we can bring. CM
Al Muñoz has many years of experience buying technology for the federal government and has held leadership roles in acquisition and information technology for some of the largest federal agencies. He has worked on bringing innovative technology and advanced contracting techniques into the government and is a frequent speaker in acquisition and leadership. Muñoz is a Fellow and recipient of the Excellence in Contracting Professionalism award.
ENDNOTES
1 The Impact of Artificial Intelligence on the Future of Workforces in the European Union and the United States of America, retrieved September 15, 2023 from https://www.whitehouse.gov/wp-content/uploads/2022/12/TTC-EC-CEA-AI-Report-12052022-1.pdf
2 The fundamental differences between automation and AI, retrieved 9/15/2023 from https://venturebeat.com/ai/the-fundamental-differences-between-automation-and-ai/
3 How AI Handles Uncertainty: An Interview With Brian Ziebart, retrieved 9/15/2023 from https://futureoflife.org/recent-news/how-ai-handles-uncertainty-brian-ziebart/
4 AI in Procurement, retrieved 9/15/2023 from https://sievo.com/resources/ai-in-procurement
5 Why generative AI needs a creative human touch, retrieved 9/16/2023 from https://mitsloan.mit.edu/ideas-made-to-matter/why-generative-ai-needs-a-creative-human-touch
6 AI, automation, and the future of work: Ten things to solve for, retrieved 9/15/2023 from https://www.mckinsey.com/featured-insights/future-of-work/ai-automation-and-the-future-of-work-ten-things-to-solve-for#part2
7 Why AI Is Never Going to Replace Subject Matter Experts and How to Choose the Right SMEs, retrieved 9/16/2023 from https://martech.zone/ai-vs-sme-generated-content/
8 How To Incorporate AI Into Your Business , retrieved 9/16/2023 from https://www.upwork.com/resources/integrate-ai-in-business
9 Three Steps to Implement AI, retrieved 9/15/2023 from https://sloanreview.mit.edu/article/three-steps-to-implement-ai/
DISCLAIMER The articles, opinions, and ideas expressed by the authors are the sole responsibility of the contributors and do not imply an opinion on the part of the officers or members of NCMA. Readers are advised that NCMA is not responsible in any way, manner, or form for these articles, opinions, and ideas.