With new AI tools being announced daily and industry and government scrambling to understand what’s next, we sat down with Michael Miguelez, CEO of OPTiMO; Bill Hombach, Director of Engineering Professional Services; Heather Hanks, Technical Principal, Analytics; CTO Scott McCarty and Chief Growth Officer Mari-Ellen Testa to learn more about the company’s AI Playbook, their advice for navigating constant change, who makes an ideal partner, and why AI for the sake of is never the answer.
Navigating a Changing Ecosystem
While the federal government is undergoing a time of rapid change and transformation, it is also true that the business of government continues. Recently that has meant that industry knows on some level what is needed but may be less certain about what specifically government is going to buy and how. “With changes to FAR, consolidation of schedules under GSA and so on we’re seeing a process change that many are still trying to understand, still watching to see how it will all roll out.”
In the face of what can only be described as unprecedented change, it is imperative that industry remain flexible and agile, able to reframe and communicate in the new language that is unfolding. In some cases that will require a reactive response, however, proactive forward movement is also attainable. “There is an opportunity to work with our clients to accelerate their processes, leveraging tools that can speed up design decisions, inform development actions and help prioritize backlogs.”
Such actions can help government clients adapt to their evolving demands but also ensure industry is keeping in lockstep with new guidance and existing requirements.
Another key to navigating changing times is staying engaged with industry to hear what others are interpreting, what they’re pursuing, and what comes next. “Events with government speakers can confirm mission direction, introduce emerging priorities, and outline next steps. They also present opportunities to connect with industry leaders to learn what they’re hearing, doing and where they are moving next.”
What is Changing Also Remains the Same
Despite change, the fundamental needs in the IT space remain the same. Government still needs to procure solutions and deliver on its missions. “Government still needs digital transformation, and modernization. They need user-friendly digital experiences for citizens and for their internal users. They still need to leverage modern technology to gain efficiency so they can do more with less.”
For industry that can mean filtering through what is being said, what has been documented, and understanding what government actually needs compared to what they are asking for. “Whether it’s an attempt to capture the latest language or a push to secure AI because there’s a perceived direction to follow or simply a repetition of past requirement language out of ease, there’s an opportunity for industry to distinguish between the stated ask and actual need, to help government clarify and really define its requirements.”
Historically more risk averse, there is a balance to be achieved between moving government forward with new things, like AI platforms and large language models, and implementation focused in the right ways and to achieve real efficiency.
An AI Playbook
Even over the past several months there has been an influx of AI tools, chat this or that…. OPTiMO set out to do something different – to develop what is essentially a playbook for AI – OPT-AI – to help clients understand where they might apply AI and provide guidance as to how they might do so ethically and securely.
“Organized around four phases – align, plan, prepare, execute – it helps assess an organization’s readiness for AI. “You want to make sure that you’re defining goals that are aligned to their mission. Then, plan by identifying what tools they have available today, or what they’re willing to invest in and assess the quality of their data to determine whether AI even makes sense… whether it can or should be applied at this point in time.”
Ensuring any AI adoption is additive – as in not being implemented for the sake of – and that its interaction would make a process more simple, not more complex is the basis behind this approach. “Ensuring AI is explainable, that a human can confirm the information is also critical. It is also important to understand that the solution built for one client, will not be the same for another. A singular tool will not meet everyone’s needs.”
Rather than stretching too far with agencies, some of the most valuable approaches will focus on meaningful efficiencies. “If we can help one government role reclaim hours occupied by mundane tasks, by applying common sense and practical forward movement—and demonstrate that these actions are scalable and translatable—we can help clients move ahead on the more important parts of their mission.”
One example of a task that can be simplified and improved through AI involves agreements or Memorandums of Understanding (MOUs) to help with the task of document management. Any area in which there is a process bottleneck, where it is possible to automate to get information or results faster is an opportunity for meaningful implementation of AI. “Procurement is another process that is time consuming. Low-code/no-code has been able to help in some cases, but AI now provides an opportunity to take that next step in improving the process. It doesn’t have to be a massive undertaking, but it can generate massive efficiency.”
The Nuts and Bolts of the Challenge
The team notes that the real challenge is not implementing a new technology, but ratherthat legacy processes may not be designed to support it. Adjusting those processes is essential to ensure success and alignment. For example, when adopting AI tools, existing review workflows may not account for the fact that outputs may contain errors or “hallucinations.” “We see with hallucinations, that some of AI produces isn’t quite right, that AI gets maybe 90% of the way there. There still need to be controls in place to evaluate and confirm the output.”
Each AI platform has its own way of handling inputs and outputs. The strategies to reduce hallucinations aren’t universal, they have to be tailored to the tool being used. Clear prompts and strong guardrails are essential. If the AI doesn’t have precise instructions, it may generate unreliable responses. Guardrails are ways of constraining the AI so it stays on-topic and avoids guessing when information is missing and AI should only be allowed to reference approved knowledge sources.
A New Way to Look at Change Management
One advantage of AI can be that the user doesn’t have to learn a new language or new tool. The user does though have to know and understand how AI has been integrated into the process.
“Having sound process around what change looks like, combined with a user centered approach that considers the user experience, is what is going to make people comfortable using an AI application. They should be able to recognize how it has improved the time they spent, the outcomes they achieved. This goes back in part to not just listening to what clients think they want when it may not resonate with users. The balance is in delivering what stakeholders want and what customers actually need and will use.”
Efficiency in the Federal Space
Noting that the goal for federal clients is intelligent automation and intelligent workflows, the goal is not just digitizing a form, but instead digitizing the form and including the human checks required to support final decisions. ”It’s not about AI making the decision, but about providing the information needed to shrink the decision process from 3 hours to 10 minutes, while maintaining the guardrails and governance that will allow people to be comfortable with their decision based on that information.”
Developing those trusted tools, and reliable responses, requires deep understanding of the agencyand topic at hand to cut the learning curve for users.
That deep understanding also realizes that large language models, now so accessible to government, may not be a solution to any of the real challenges agencies face, particularly if they are being bought without knowing why or how they might be used.
“We’re sitting down with some of the smartest people in the industry and we’re figuring it out internally. We’re sharing ideas and we’re building tools and we’re building workflows and trying to apply it to customer needs and we’re coming up with these great solutions in these great frameworks. We still see the same things in procurement where they’re buying AI for the sake of buying AI, but they’re not buying a way to implement AI so then they have this tool they have to figure out a use for.”
Understanding whether AI is needed, whether that may make a process overly complex, seeking a better solution should still be the name of the game for the vendor community, should still be how industry works with federal clients to recognize where they can gain efficiencies and the best way to do so.
The Role of Procurement
“From a procurement perspective, it’s the difference between a performance work statement and a statement of work. The performance work statement is what do I need? What do I want? What are my objectives? Where are my pain points? What do I want to accomplish at the end of the day. An SOW says I want this, including AI. Procurement really needs to think about and address pain points, objectives, and what they want to see out of the effort. And then they need to let industry help them in their responses so they get the right solution.”
Often before AI is introduced, the groundwork needs to be laid, ensuring the data is ready, considering adoption processes and other aspects that need to be navigated. “Part of our playbook is helping organizations understand whether they are ready, or what they need to be ready.”
“We can always find quick wins, help show them the way so they start to grasp what is possible, can start to see the benefits. It comes down to listening to their processes, asking about challenges.”
“Leaders should ask their teams the five tasks they spend the most time on every week. Then they can go to their trusted vendors to say, what do you think we could do, is there something we can implement to solve this challenge? The way AI works there are likely efficiencies to be gained for every one of those tasks. If they can do that for every member of their team, those minutes equate to massive improvements and time savers.”
Noting that massive efficiency doesn’t necessarily mean massive procurements, there is opportunity for smaller projects that can demonstrate and then be scaled to much broader adoption. “Measuring success for any project can be easily defined. Is it reducing the number of steps or time to do something? Is it reducing the number of clicks? Is it increasing your ratings? What does success look like?”
The Partner Lens
Having previously noted its efforts working with others in industry to find the best solutions, to understand the way things will be moving forward, OPTiMO says it is all about partnerships, about working with others in the vendor community. “Small businesses are at a disadvantage to larges with more limited resources who may be able to invest in more that may not come to fruition. Identifying those who have needed capabilities and customer intimacy, those who understand that teaming is the lifeblood in this environment, and that culture plays a big role, both in internal success and client delivery, is our goal.”
“For OPTiMO, the way we treat each other, show up for our customers, treat our partners is ingrained in who we are and is non-negotiable when it comes to who we will choose to partner with.”
