A terrific team of judges helped draw a large slate of entries for this year’s NCMA Innovation in Contracting award. They vetted each one closely. Two were so strong that the panel couldn’t narrow down to a single winner. That’s why we are shining the spotlight in this issue on Tradewind and TryAI, both created by the Defense Department Chief Digital and AI Officer (CDAO) Joint Artificial Intelligence Center (JAIC).
This column has been home to two articles by Andrew Bowne in 2021 and 2022 based on his deep study of what attracts and repels innovative artificial intelligence (AI) firms about the Department of Defense (DoD) market. So, it’s fitting that the Innovation Award goes to two initiatives aimed at making it easier to attract and contract with them.
Consonant with Bowne’s finding that DoD’s treatment of intellectual property (IP) rights is a turn-off for AI firms, Tradewind is working on IP guidance and specific language for AI deals and TryAI is using modular contracts to help mitigate risk. Bowne also found that AI firms prefer commercial-style contracting to Federal Acquisition Regulation-based. Both award winners are using practices outside the FAR to help defense programs try and buy AI. Tradewind applies other transaction authority, while TryAI wields the commercial solutions opening.
We welcome Carol Tisone to Innovations as the author of this month’s column. She is the former Defense Acquisition University Dean of the Mid-Atlantic Region and will appear more often in Contract Management in coming months. We are privileged to have her.
Thank You 2022 Innovation in
Contracting Award Judges
• Ben McMartin, Public Spend Forum
• Michelle Currier, NCMA
• Polly Hall, Department of Homeland Security
• Trish Martinelli, Defense Entrepreneurs Forum
• Victor Deal, HII
Smarter AI Acquisition - Is the Focus for 2022 NCMA Innovation in Contracting Award Winners
DoD programs Tradewind and TryAI add expertise and free trials to the AI-buying toolbox.
BY CAROL TISONE
It isn’t always easy to imagine how artificial intelligence (AI) can help solve a program challenge. Few program or contracting professionals have a comprehensive understanding of the AI market or participating companies.
Most AI companies are commercially focused and unfamiliar and uncomfortable with federal buying rules and practices. What’s more, it can be costly in terms of training data and dollars to buy AI solutions without trying them out first.
Enter Tradewind1 and TryAI,2 this year’s winners of the NCMA Innovation in Contracting Award. Both are creations of the Defense Department Chief Digital and AI Officer (CDAO) Joint Artificial Intelligence Center (JAIC).
The CDAO/JAIC Tradewind initiative is an implementation vehicle for applying and assessing innovative contracting processes for fast delivery of AI capabilities. Tradewind applies new policies without traditional red tape and quickly assesses proposed solutions for effectiveness and adjusts them in real time as part of the procurement process.
TryAI, a commercial solutions opening (CSO) model, allows customers to test innovative technologies before they buy them.
The NCMA Innovation in Contracting Award recognizes successful applications of innovative or leading-edge contract management practices in ways not previously demonstrated to achieve mission or organizational goals and results. Submissions must describe the business value of a project; the degree to which it is in use, how it exemplifies leading practices other organizations can adopt; and its use of innovative concepts, strategies, or practices.
Bridging AI Expertise Gaps
Tradewind is designed to provide an ecosystem that will quickly transition a concept from an idea to a working project. It tests the proof of concept through application and analysis of real results under an array of solicitation methods and other transaction authority (OTA).
The Tradewind ecosystem engages top AI innovators in industry, academia, nonprofits and government. Using agile methodologies to monitor and assess work output, the CDAO/JAIC team quickly homes in on what works and what doesn’t during contract execution and adjusts accordingly. With seven project awards in nine months at this writing, Tradewind has achieved significant efficiencies and removed traditional barriers while delivering AI capabilities. Prototype agreements awarded in 30 to 60 days enable the DoD to quickly, repeatedly, and ethically operationalize AI technologies.
While Tradewind appears to be a standard process (See Figure 1), it’s the execution that supports the innovation.
Tradewind uses a data-centric acquisition infrastructure that includes an online portal featuring single-sign-on, tech scouting, market research, evaluation tools, and the development of best practices. The initiative offers a means to bridge AI expertise gaps and improve the contract decision-making process.
For example, when seeking solutions to address an AI data integration problem for combatant commanders, Tradewind formed an AI technical advisory board of industry experts and academics, in just five days, to support the solicitation process. Experts brought experience in intelligent model-based reasoning and data analytics as well as AI/ML (machine learning), high-performance parallel and distributed computing, and cloud computing.
The board participated in white paper reviews and one-on-ones with vendors to independently validate findings throughout evaluation. The team winnowed 74 white papers to 11 solution briefs to four 2-hour one-on-one discussions and demonstrations, making a final selection in three weeks. Access to the right expertise at the right time was key to swift decision making.
Connecting Puzzle Pieces
Tradewind is an “enablement environment” aimed at decreasing the time from the “point of ideation of an emerging technology to prototype and transition” by developing a variety of strategies using data sets and demonstration environments, explains Bonnie Evangelista, senior procurement analyst and Tradewind manager. Tradewind is focused on overcoming traditional barriers, perceptions, and culture to pursue cutting-edge AI capabilities, she adds. The goal is to take acquisition and contracting to a new level through innovation and collaboration.
“The name of the game is connecting puzzle pieces so it’s a win-win for everyone – industry, government, and the warfighter,” Evangelista says.
The team is trying to create a different culture by bringing industry, academia, government, and nonprofits to the table to create an active ecosystem. “Greater communication and collaboration from working together will help us realize what we didn’t know before,” says Evangelista.
Tradewind’s non-Federal Acquisition Regulation contracting environment provides much needed flexibility in pursuing cutting-edge AI technologies. The CDAO currently uses the Army Contracting Center in Rock Island, Illinois, as its contracting office while waiting for its contracting delegation to be finalized. Evangelista hopes to provide rotation or embed opportunities for others to experience the Tradewind environment to “change hearts and minds.”
“We want to open up the aperture for what people think is possible,” Evangelista says. “Folks are checklist- and process-oriented and may not understand that more is possible.” By disrupting what people believe acquisition and contracting are supposed to be from a process and business model perspective, Evangelista hopes to remove barriers, try new models, and help groom an AI-savvy workforce.
Try Before You Buy
TryAI is a commercial solutions opening (CSO) model that allows DoD customers to try innovative technologies before they buy them. TryAI combines three innovative concepts to create one flexible contract management solution.
The first is the CSO; which is a merit-based, general solicitation designed to allow DoD technical experts to use their best judgment in deciding which capabilities might help solve the defined problem. Technical evaluation is streamlined, and negotiations are rapid.
The second is OT authority, which allows the team to draft five-page demonstration agreements that limit the workload for DoD, vendors, and technical leads.
The third is modular contracting, which breaks up the acquisition to isolate failure and mitigate project risk. Modular contracting provides competitive leverage while negotiating pricing, intellectual property, and other terms.
Demonstrations are a significant factor in realizing innovation. They occur in DoD enclaves – separate secure networks – and on DoD data, and the team does not always pay the vendor for the 90-to-365-day demonstrations. To date, the team has conducted a range of demonstrations including, for example, computer vision enhancement and source selection automation.
The approach mitigates cost overruns and execution risk by proving a technology can perform as proposed before acquisition. No more paying for software and not seeing a return on the investment. The team can quickly and effectively work through white papers and negotiate in real time the parameters of a prototype and demonstration.
Keith Gibson, acquisition analyst with CDAO/JAIC, says the idea is to “fail fast – take execution risk down to zero because the demonstration is not in [a commercial] environment [where] you have to migrate it, [and] because you’re doing it in real time. If it’s a hit, the benefit is at the end of demonstration; we can immediately award a follow-on on contract.”
The CDAO/JAIC team is seeing positive responses from the DoD technical leads as they see promise in the demonstrated capabilities. The enhanced discussions within the ecosystem enable extensive market research, always a challenge for innovative technologies.
Vendors like the streamlined submission criteria and demonstration agreements. They also appreciate the feedback on their products via reports at the conclusion of demonstrations. The hope is that vendors’ positive experiences with TryAI will encourage them to engage with DoD more regularly to improve government and industry interactions.
Evangelista and Gibson agree that finding people willing to do something new has been an obstacle. Through discussion, demonstration, and the positive outcomes achieved so far, they hope to help acquisition professionals challenge traditional customs and preferences for what is known and comfortable and try something different. Education and awareness are key to creating comfort with the unknown.
JAIC is in the early stages of creating an AI Acquisition Experiential Learning Ecosystem through Tradewind to pilot using industry or academic teams to guide DoD acquisition professionals through agile AI projects. The model will provide training-with-industry opportunities and create a partnership model for AI training and education.
Tradewind and TryAI are available for use within DoD through teaming with the CDAO office. Ideally, the TryAI CSO and Tradewind ecosystem will reduce barriers for nontraditional contractors to provide DoD innovative commercial items, technologies, and services.
If a DoD customer has a data set, a demonstration environment, and technical experts, the CDAO team is there to help set up a demonstration. If the demonstration proves a technology is useful, because CSO is merit-based, a follow-on contract can be awarded quickly.
Tradewind has drafted and published guides for better planning and negotiation, as well as recommending intellectual property language. These publications resulted from DoD-industry interaction, demonstrating how the collaborative environment creates impactful contributions.
As Tradewind continues to design and build practices, it empowers DoD to implement rapid procurement and delivery of AI capabilities, embodying the Tradewind “rapid done right” principles.
JDAO hopes to expand the Tradewind ecosystem across the military services. CM
Carol Tisone spent 23 years in contracting and acquisition with the Naval Air Systems Command. She joined the Defense Acquisition University in 2008, serving first as contracting faculty and Department Chair, then moving on to Academic Dean, and Dean for the Mid-Atlantic region until 2020.