The Buzz with ACT-IAC
The Buzz with ACT-IAC
ICYMI: AI Revolution in the Acquisition Life Cycle AI Acquisition Forum- 2025
This panel discussion, moderated by Joanie Newhart, Associate Administrator for Acquisition Workforce Programs at the Office of Management and Budget, delves into the impact of AI on the acquisition lifecycle in government agencies. The panel includes Bonnie Evangelista, Jonathan Mostowski, and Kaprice Tucker, who share their diverse experiences and perspectives. Topics cover AI applications in defense and civilian sectors, overcoming fear and resistance, optimizing processes, and the future of AI in acquisition. They also discuss the role of leadership in promoting innovation and the importance of exposure to AI to reduce fear and improve adoption.
Subscribe on your favorite podcast platform to never miss an episode! For more from ACT-IAC, follow us on LinkedIn or visit http://www.actiac.org.
Learn more about membership at https://www.actiac.org/join.
Donate to ACT-IAC at https://actiac.org/donate.
Intro/Outro Music: See a Brighter Day/Gloria Tells
Courtesy of Epidemic Sound
(Episodes 1-159: Intro/Outro Music: Focal Point/Young Community
Courtesy of Epidemic Sound)
Soraya Correa: [00:00:00] We have an exciting panel that's gonna talk about one of my favorite topics, which is the AI revolution in the acquisition lifecycle. Um, to moderate this panel, we have Joni Newhart, one of our chairs, associate Administrator for Acquisition Workforce Programs at the Office of Management and Budget. And she has several panelists that'll be joining her.
Soraya Correa: Hopefully they're all here. Uh, Bonnie Evangelista, senior acquisition official. Jonathan Moki, president of Agile Acquisitions, LLC Caprice Tucker, associate Director for Acquisitions at the US Department of Interior. And Kristin Wilson, acting Associate Administrator of Office Management and Budget. So welcome to the stage and take it away on one of my favorite topics.
Joanie Newhart: Thank you so much Sariah. Very glad to be up here again. And, uh, a little program note, Kristen Wilson couldn't make it so. Um, you'll have to meet her another time. She's great. But we have three stellar panelists. So let's just get to the introductions. [00:01:00] Could you each, um, talk about your background, your current role, and how you interface with this new monster ai?
Joanie Newhart: Okay,
Bonnie Evangelista: great. I'll start. My name is Bonnie Evangelista. I have supported federal and defense agencies for 17 years. Uh, I've been in the Department of Defense for. Almost, oh my God, I've lost count seven or eight years now, and I landed in the AI space 'cause I was part of the Joint AI Center, which has now turned into the Chief Digital and AI office.
Bonnie Evangelista: But more recently I've kind of transitioned from them. I'm advising on the revolutionary Far Overhaul team and helping with the far rewrites and whatnot. So I have some breadth and depth in this space. Uh, I've done a lot of experimentation, a lot of Zach's comments resonated really well in terms of acquisition, not just acquisition of ai, but AI in acquisition.
Bonnie Evangelista: It's a topic that's near and dear to my heart for that. So over to you.
Jonathan Mostowki: Uh, very happy to be here. Thank you. Uh, Jonathan [00:02:00] Salki. So I started my career as a contracting officer at NGA. Spent about 13 years there, and, uh, kind of was forced into a situation where I had to solve how to buy Agile. Software development services and that kind of opened a series of doors of innovation opportunities, led me over to the US Digital Service and defense digital Service in the early days.
Jonathan Mostowki: Um, and then I left to go help do a tech startup that was focused on AI solutions to DOD National Security Challenges. Uh, I currently have my own, uh, consulting company where I work with government agencies. Uh, directly kind of in a digital service sort of capacity, as well as companies that are trying to, uh, work with the federal government.
Jonathan Mostowki: And on both sides, AI is front and center. Uh, you know, from the startup perspective, it was really, you know. Using AI for kinetic activities and decisions. So very high consequences, very early on in this AI journey. And now it's a kind of persistent small [00:03:00] businesses. How do we compete by using AI and and government?
Jonathan Mostowki: How do we keep up with demand, uh, and in some cases shrinking workforces by leveraging ai. So happy to talk about this today.
Kaprice Tucker: Hello everyone. I'm Caprice Tucker. I'm from the Department of Interior. Associate director for acquisitions. I actually started my career as a contract specialist, uh, back at nasa.
Kaprice Tucker: Spent over 20 years there between, uh, being an 1102 in the contract space and also in the Office of Chief Counsel. So I've bounced back and forth between legal and. Contracts over my career. I'm at the Department of Interior now, as I mentioned, and I think, um, a little different in the sense of, I feel like I'm just the champion to get innovation integrated in our area so that we can do more with less.
Kaprice Tucker: So, uh, right now it's become even more prevalent, um, that we need it as [00:04:00] things change, but really looking, uh, for ways to improve the. Efficiency of our contracting workforce and also serve as our acquisition innovation advocate. Shout out, um, to that group across the department. So I'm happy to be here.
Joanie Newhart: Great. So you can see we have three different perspectives up here. So this is gonna be a really good panel. I'm really happy to be part of this. Um, so as somebody mentioned, as everybody mentioned, this is happening at the speed of light, like it's happening so fast. We at O-F-P-P-I know. I don't think we know everything that's going on, of course, but we want to get more of a handle on it so that we can ensure there's a lot of sharing going on, because we don't want all the agencies to figure out everything all by themselves.
Joanie Newhart: But, so from your perspectives, what do you see going on in the AI and acquisition space? Um, you wanna start Bonnie?
Bonnie Evangelista: Sure. Uh, and my observations, uh, even though I've spent a lot of [00:05:00] time in the Department of Defense, I am getting a little bit of inkling on the civilian side. So it's coming from both of those perspectives as well.
Bonnie Evangelista: There. The fear thing that Zach talked about is real, number one. Um, and that is not easily overcome. And even when you have. An awesome solution. Or maybe it's not even a tool, maybe it's just a methodology that could encourage more use in the future. There is still massive fear, not just of the thing, but someone's gonna say no.
Bonnie Evangelista: So that fear culture that no culture is not easy to navigate. And then the other thing I'm seeing is there, I thought the last question that Sariah read. Is the million dollar question, because even though I am seeing a lot more AI use in the DOD, we have nipper, GPT, and that was an experiment that Army, or sorry, air Force Research [00:06:00] Lab initiated to give instant access like Zach was talking about, give users just something they can play with and then they can look at and see how people are using it and what is viable or effective for possible scaling and into a production environment.
Bonnie Evangelista: Army's reaction to that was we're not gonna use it. So they were able to use it for a while and then they, they said there was a security risk and they shut it down and they bought some other tool that, I don't know if it's being used or not, but I guess like that's just an observation to, for everyone to look at.
Bonnie Evangelista: The other services are using it. One service is like, yeah, no thanks. Not for us. And what do you do with that? But, um, I am seeing more of these general purpose. LLM chats, interfaces being used, uh, I'm seeing that at GSA as well. They have their own, um, the getting to production thing. I'm not seeing, I'm seeing a lot of experimentation, a lot of things [00:07:00] being done under the guise of experimentation, and I'm not seeing us maybe finding either rapid or more, um, quick.
Bonnie Evangelista: Pipelines into something that's in a production environment that we can say, this is the thing, use it for this use case and do better, do more. I'm also seeing a lot of, um, eagerness from industry, like they have the cutting edge and. Uh, we, they were, we were just hearing about agentic ai. I've been hearing about agentic AI for over a year, and if you, so if you talk to industry like this is not new.
Bonnie Evangelista: Um, I've been hearing from industry, like, whether you agree with this or not, I'm just like sharing observations. Um, data hygiene and labeled data, things like that, that's like obsolete technology. Like there is technology now that can go beyond that and still. Do what? Get the answer you're looking for. Do what you need it to do [00:08:00] and, and things like that.
Bonnie Evangelista: So like the answers are out there, just how are like, so I'm not seeing. Us really optimize that in a timely fashion. Like, like are we actually exploring those things now versus like a year after industry figures it out and figures out how to deploy open AI does their version of that and we finally get on board with it.
Bonnie Evangelista: Um, and then the last thing I'll share, I'm so sorry. I, and is, I, I do think CDAO did something very interesting and they recently awarded. Contracts to all of the foundational, I'm sorry, frontier model companies. And they basically said, go to the combatant commands and just tell us how to do this better.
Bonnie Evangelista: And I think there is not enough of that. So whoever is cutting edge, giving them some kind of box, you know, whatever that box is, they gave him a box. They gave him, they only gave him a little bit of money. They gave him a mission or, or an objective, and they said, tell us. [00:09:00] What right could look like or what the art of the possible is.
Bonnie Evangelista: I'm not seeing enough of that kind of experimentation because we're so far behind. I feel like we're still trying to catch up with like, Hey, we have these use cases here. How do we just get them squared away?
Joanie Newhart: Gosh, thank you that lot. There's a lot there. Sorry. So department of Interior, um, my government co-lead, who you'll see later, Andrea Brandon is there and I met her a few years ago.
Joanie Newhart: She's the most leading edge emerging technology. One of the most leading edge emerging technology people I know. And I'm just riveted hearing her a lot, like I'm riveted with these guys. Like this is all new stuff for me. And I'm just like, what? What? We could do that. We could do that. So Caprice, um, you're at Department of Interior, so what kind of use cases and pilots?
Joanie Newhart: I are you guys on the cutting edge? I feel like you might be.
Kaprice Tucker: Well, I would say, um, we have definitely been using, uh, ai. We've, in acquisition specifically, we started with [00:10:00] RPA and, um, automated some of our processes. We, um, I think most of our success so far has been with our bot to do closeout. We've done a hundred.
Kaprice Tucker: Thousand closeouts, I believe in the last couple years that has taken off a lot of pressure from our contracting officers. That's in an area where, uh, we don't have a lot of resources, so it's kind of like we'll get to it when we can and we never get to it. So that has been very helpful. Uh, and then.
Kaprice Tucker: Spinning up from there. We've done, across the department in general, we have our own DOI chat, GPT, which is behind our firewall. We've done many, I would say we're still in the use cases stage where all across the department, the last I've heard we've had about a hundred different uses. Um, but in acquisition specifically, we work in [00:11:00] pockets.
Kaprice Tucker: So we find the people that are excited to work. Start there. Uh, they start using the, um, D-O-I-G-P-T all across the acquisition lifecycle. And then we try to share those successes to get people, um, kind of like, um, what we were talking about. It's. Trying to get rid of the fear or at least allowing people to proceed in spite of the fear.
Kaprice Tucker: And so, um, we've seen some cases where we've just done brown bags or training in how to use our DOI chat, GPT in acquisitions, showing some, uh, demos on how you can use it, whether it is asking the. Uh, or with a prompt, uh, for a learning model just to say, you know, how can I analyze the statement of work or how can we look at requirements?
Kaprice Tucker: And, um, just kind of all across the board. Training has been a big one where [00:12:00] we've asked to break down some complicated far, uh, section. Not that the far is ever complicated, right? Um, but. Uh, seeing help in those areas where people are kind of getting just a little bit used to it.
Joanie Newhart: That's great. Caprice.
Joanie Newhart: Thanks. So Jonathan, we're hearing industries way ahead of us. You're now industry. Tell us what you're doing.
Jonathan Mostowki: Uh, sure. Uh, and actually I'm just gonna take one step back though. Uh, and I'm gonna temper what I, I'm about to say with. Zach's point of planting lots of flowers and letting them grow. But, um, I would say the government is doing what the government always does.
Jonathan Mostowki: They are building a thousand discreet solutions to the same problem. Um, you know. Department of Interior Chat, which I hadn't heard of before, but I'm not surprised. State Department Chat. There's a number of variations at DOD, you have ACT Bot, you have, uh, as SAGE as a commercial solution, you've got the Air Force's version.
Jonathan Mostowki: Uh, I mean, it's, we're [00:13:00] just o we're, we're building and we're building and we're building and everybody's building to the same place on the same foundational solutions at the same time where it's saying to industry, we want commercial, we want you to build it. Use venture capital we want to invest in, in that direction, but we're doing it all to as well.
Jonathan Mostowki: And so I, I think there's, uh, that's creating a little bit of dysfunction, uh, on the industry. Uh, well, and I, and I'll just say, you know, a lot of what we're trying to do is also a very government activity, which is over complicated. Uh, there are capabilities today. I mean, I, I trained an ACT bot on my own material and I use it to help me, you know, answer my own questions or questions for clients because, you know, it's just faster, it's easier.
Jonathan Mostowki: So, you know, when you get to like acquisition source, selection, sensitive type stuff, yeah. That bar is really high. You're starting to get kind of to the same areas of like military activities. You need a human in the loop. When it comes to just generating requirement [00:14:00] documents, that technology exists, that technology is basically free.
Jonathan Mostowki: And the only thing the government should be worried about is securing data. Uh, on the industry side, what I'm seeing is what you, a similar activity, but it's what you would hope to see in industry, which is the industrial Revolution. Race to the front with the best solutions, the best metrics, the best capabilities.
Jonathan Mostowki: That's what we want, and that's who should be doing this. So we have to decide, I think as a government, do we want to pay for custom solutions over and over and over again? Or do we want to invest in industry to develop solutions that can be configured to each agency's needs? And you know, I think that's the question to be answered, but, um, I, I get contacted by venture capital firms regularly to evaluate a new commercial possible solution for acquisitions.
Jonathan Mostowki: But what's gonna happen if, uh, you have commercial capabilities for writing proposals? You have government capabilities for drafting RFPs and evaluating or evaluating proposals, right? All of a sudden [00:15:00] it's like, who's actually asking for something? Who's actually bidding? Like, what are we doing here? Uh, and so we, and if everybody's doing it differently, that's even worse, right?
Jonathan Mostowki: So we really have to come to a collective mindset on what is the solution we're looking for here. I deal with small businesses all the time. They're like 10 people. Like we use AI to submit all of our proposals. We win a lot, and we use AI to write the code. We're not hiring any more engineers. It's wild.
Bonnie Evangelista: And there's technology now that like you can hire an AI team. Yep. Rather than expensive consultants. And so that barrier to entry is lowering, which some would argue is a good thing for small businesses. But the problem, it is, uh. It's not on the government side. We're not handled to meet, we're not postured to handle that scaling of we're gonna get more and more and more.
Bonnie Evangelista: Well, how are we gonna evaluate and then do our part. So we're gonna be the delay now unless [00:16:00] we actually take this seriously. And then you mentioned like. Kind of an example of, I think what you're talking about with like, we should let industry lead this. They are leading it and we're still catching up.
Bonnie Evangelista: I'm still hearing things in the government where, where people say like, the solution is, I just want the TurboTax of acquisition. I'm like, that is old technology. Why do we want the TurboTax of acquisition? Like let them build. The matrix, you know, or whatever that, you know, is, is cutting edge and that, that's like kind of the mindset we're, I think we're trying to tease out here is like that's what people are navigating or battling right now.
Joanie Newhart: Wow. It's way more complex out there than I thought. I think I'm gonna need a drink after this, but, so a, i in acquisition, it seems like a, a very good marriage, right? Um, a I would help in so many spots in the acquisition process, I can see it, but as I go back to my contracting officer [00:17:00] days, I feel like if I had ai, I would be like.
Joanie Newhart: Oh, that's great. But I don't trust it. I have to go back and check everything, which causes me, you know, 10 times more work. I'm, no, that's not the right mindset. How do we get past that kind of stuff and how do we move the, the dial forward here? You
Bonnie Evangelista: know what's funny when you, so I was also a contracting officer and like when this, all this AI stuff happened, I was excited.
Bonnie Evangelista: I was. Holy crap, I can go faster or I can communicate better with, I was part of ACT Bott, like one of the biggest appeals of Building Act, Bott and this, and we built that before chat. GBT was, or we started playing with it before chat. OpenAI released the floodgates with chat. GPTI had no idea this technology existed and they showed me an MVP.
Bonnie Evangelista: They took a single title and it started generating three paragraphs of a problem statement. And I was, and I was. My mind was blown away. So I was so excited because I saw the potential of how this could change, like how we do [00:18:00] business. Um, and maybe that's so the, there's different personas here, but there you have a, it's a good question.
Bonnie Evangelista: Like, how do we deal with that? Because everyone has a different lens. Um, I, I always took the approach of, um, most people. You know, don't wanna be drenched in paperwork. So if I could help them see, oh, like you won't have to spend hours doing task X. You could spend maybe 30 minutes doing Task X, you know, and that usually lended to a more productive conversation.
Bonnie Evangelista: That was that I'll offer that.
Kaprice Tucker: I'll say, um, for me, I think it's important to use your partners. You know, whether it is our, we have a business integration office. They are our technical arm that will work with us to implement any kind of technologies across our space, the OCIOs office, um, and get the people that, you know, are driving the train and hooked [00:19:00] together so that.
Kaprice Tucker: You have someone that's advocating for you so that you can, you know, break it down into real life examples. And then, um, I feel like my role is to just make sure I'm advocating for the acquisition workforce so that they can see those, oh my, you mean to tell me that I don't have to spend all this time in something that's not important.
Kaprice Tucker: And then assure them that they will still have. Input in the process, they will, their expertise, their actual subject matter expertise will still be needed despite the use of the ai.
Jonathan Mostowki: Yeah, I, I think, I think we heard a lot of it in the keynote introduction here. Uh, you know, it's exposure, it's getting people.
Jonathan Mostowki: Used to what it can do and can't do. Uh, there's a human in the loop. It's just a question of where you put that human. And as you get more confidence, you can adjust the placement. Um, but to be honest, I mean, I was a contracting officer for a very long [00:20:00] time. We used Prism. The clauses were never right. You know, and that wasn't ai like you had to go through.
Jonathan Mostowki: As a subject matter expert and make sure the right clauses were there and, and they usually weren't all there. And that was just part of the job. And so, again, like I, I think we're, we're kind of at this like dichotomy of like, we can see the future. Of AI is gonna automate all this stuff and we can just sit back and, you know, let that happen.
Jonathan Mostowki: But we're not actually there right now. Like there's actually a long way to go. And so there's al, there's not always going to be a human in the loop necessarily, but there's always, in the near term, going to need a human to verify the data is correct. But the, the metrics and the accuracy of the models will help us determine how much and where that human has to get involved.
Jonathan Mostowki: And, and honestly, you know. What is the risk we're talking about, right? We're gonna use AI to produce our RFP. Maybe it included the wrong clause. It's okay. 'cause industry's [00:21:00] gonna run it through their AI and the AI's gonna flag it. I mean, I'm not kidding. There's already a solution right now that is AI based that validates whether the right clauses are used.
Jonathan Mostowki: And I have companies who hire me all the time to say, can you check these clauses and make sure they're right? And I'm like, yeah. I mean, come on, let's go through the commercial clauses and see, uh, you know, I mean, so this, I, I think this is a solvable thing. It's just we have to step to it and stop trying to be way over here in the future when, at least for government's sake, we're way back here.
Jonathan Mostowki: Right? Yeah. It's all, I like
Bonnie Evangelista: that because, like Joni, when you were a contracting officer, what was the most painful part of your job? Do you remember?
Joanie Newhart: I think the clauses might have clauses. Okay. We
Bonnie Evangelista: didn't have it automated. We should pick the most painful parts. Of whatever function we're trying to solve in acquisition, in contracting.
Bonnie Evangelista: Uh, I don't, I remember reading the clause matrix that, you're right. It was very painful. I didn't, we used a different version of Prism and I didn't get any suggested [00:22:00] clauses, so I just had to do it myself. Um, but whatever that is today, um, generating documents and, and, and, uh. I would go a step further.
Bonnie Evangelista: Sometimes it's not even generating the documents, it's just getting consensus on the document that is a painful part of the process. Like pick a pain point and be like, I think I have a, you know, we have a way to experiment around this use case and give you, uh, the person who might be skeptical of the tool, but like, what if we solve that pain point?
Bonnie Evangelista: And I remember talking to the chief learning officer, the first chief learning officer at the CIA, and he talked about that a lot because his job was just to like go. Be nosy into other people's problems. But he said most of the time when you offered to like solve a pain point, like mo, many people were very open to not just conversation, but like letting you in to their world and exposing them to something different.
Jonathan Mostowki: So real, real quick, Jon. So I'll just say, you know, talking about what's the pain point? It's interesting. So I, I helped create the [00:23:00] dite program with, with Joni and Tracy Walker and Mike Palmer's out there as well. And, uh. I still continue to help facilitate it and we do the capstone projects. And now the capstone project is basically like teams of, of the students or the cohort, uh, coming up with a real problem and sort of working as a product owner to develop a solution.
Jonathan Mostowki: And I can't tell you how many times it is replacing the checklist of what goes in the contract file. It's never make clauses easier and, and maybe 'cause they're just not even thinking that big. Maybe that is the big problem. But everybody, how do we improve the market research? How do we improve how we engage with industry and how do we automate a checklist so it's the right checklist for the right type of acquisition?
Jonathan Mostowki: Like I don't even think you need AI to do that, right? Like, but AI can sure as hell make that. Easier and faster. So it like, that's a perfect example of like a really, really solvable solution. So, you know, the whole like TurboTax for, [00:24:00] for contracting, I think what people are trying to say is an automated system, which TurboTax is and um, that does everything for you.
Jonathan Mostowki: And, and that's what I mean we're overcomplicating it. Pick the smallest, most important, this is agile, right? This is iterative development. Pick the smallest, most important problem that's closest to the boat, solve that really, really, really well. Then people who are like, I don't know what AI is, they're gonna be like, that's cool and I trust it and it's really not that risky.
Jonathan Mostowki: If it gave me the wrong checklist, fine. I missed one thing. You know it. That's a really solvable problem that gets people comfortable.
Bonnie Evangelista: And in addition to that, though, someone, 'cause you're right, like there, there are problems today that there's plenty of technology that could aid. Now, like right now, I do think someone needs to be thinking about, I don't think our status quo processes are the future.
Bonnie Evangelista: Right. And someone should be thinking about that. That's what, you know, I, I do think there is a suggestion there to, um, to imagine and not just [00:25:00] automate the processes that the administration is calling slow and not innovative and not giving us the results we want. So someone should be thinking about like.
Bonnie Evangelista: What does this look like? If we put all the rules to the side for a second and how can we like just get this back to first principles or whatever. You know, the administration calls it common sense, so
Jonathan Mostowki: doing the wrong thing faster is an innovation.
Kaprice Tucker: I, I'll just say from a the leadership perspective, it is giving your folks tap cover to try yes.
Kaprice Tucker: And you know, whether it is just being innovative and taking risk and being okay with accepting some of the risks. So it takes away some of that fear when they're worried about, oh, if I use the wrong checklist, am I gonna get fired? Right.
Joanie Newhart: Gosh, that makes me think of our Mc, SORAY Correa at D Doubt. S She started the the pill and she gave people top cover.
Joanie Newhart: And what we saw from that has been amazing in the years that it's [00:26:00] been around. Now it's, it's. Thank you. Now it's grown. And so the, the last keynote's gonna be Jeff from NASA who, uh, is in charge of their, um, procurement innovation lab. So I, I like that model a lot. I've seen, we're familiar with it. Now. I think the government, what, what do we need to do?
Joanie Newhart: Is that the same model we need to use here? And what I feel like at OFPP? So now this is my panel. Um, we're not doing enough. Like I feel like we're watching things. And maybe we could help somehow, um, agency leaders should help. What, what's your take on that? Uh, go ahead.
Kaprice Tucker: I'll just say, to go back to, uh, SORAY, I think one of the things is digging down deep and letting the people that do the real work fix the problems.
Kaprice Tucker: At least I Amen. I remember getting that from the pill and all of her training is, you know. I'm not writing contracts day by [00:27:00] day, but I, I do have smart people access to the smart people and letting them, they are already talented, letting them know that they're engaged and we are respecting their expertise and letting them fix the problem and providing a back to providing that top cover.
Bonnie Evangelista: Yeah. I have a, a love hate relationship with the, not just the pill, but there are groups like the pill. All over the federal government. Um, in DOD you have the Defense Innovation Unit, and then you have other innovation hubs within the department that are separate and compartmentalized. And, and I'm seeing the growth that you mentioned, NASA and I think commerce has their own version of the pill, and I I love it because you're, you're giving permission for people to imagine or play or trailblaze.
Bonnie Evangelista: And the problem I see though is it offloads the imagination from everybody else to those groups and it doesn't necessarily fix some of the cultural things that Zach was talking about. So [00:28:00] how do you get that contract specialist, that contracting officer, whoever is closest to the problem, how do you actually encourage them to generate the idea and feel confident enough to just do something different, whether it's change some language in their solicitation.
Bonnie Evangelista: Or try a different methodology or like, I, I used to, Joni knows this, like I, I did a lot of work implementing a novel methodology. There was no ai, absolutely no AI in this. It was just like doing business differently in the department. We leveraged underutilized tools like OTAs, other transaction authorities, commercial solutions, opening procedures, and then we used some familiar tools like broad agency announcement procedures and we.
Bonnie Evangelista: Smooshed it all into one general solicitation. And I was like, maximum flexibility for the CEO or Cs. And I thought this was the best idea ever. I'm like, people are gonna like use this like hot cakes 'cause I'm gonna speed up their timeline. And the top two answers I got, 'cause I trained over 500 people in the department [00:29:00] on how to use AI and how to buy ai.
Bonnie Evangelista: And the top two answers when I said, what's stopping you from using this methodology, not even an AI tool. And it was, my boss will say no and legal will say no. So back to our fear culture, but also like again, how so, so they're not even willing to try it. So how do you get something that change culture, that fear culture, like whatever it is like has to be rooted out in addition to having some like really bright organizations like the pill.
Bonnie Evangelista: Uh, but they can't be the be all end all. There has to be that like some kind of, um. Not exclusive to that organization. It has to be like, permeated through the entire organization that like, hey, there's this group that could help you, but you know, you can do it on your own if you are, like, if you've got good ideas.
Bonnie Evangelista: And that's not, I don't see enough of that being encouraged.
Jonathan Mostowki: Yeah. I, I think this is a, a really, [00:30:00] uh, interesting time. Right. Uh. I, I have a similar feeling. I, I see all these organizations of innovation and one step removed. People don't have a clue what to do. You step right outside the TAC in the va, they've never heard of using ER phase three.
Jonathan Mostowki: You step one step outside of DIU or one step out of F works and they have no idea how to do these things. So it's, it's great. There's a place to go for innovation, but if that place can't solve all of that organization's problems, you've basically just created a lot of frustration. 'cause they're like, I know if I was working with them, I could do it, and you're telling me you don't even know what I'm talking about.
Jonathan Mostowki: Right. So that, that's really challenging. So what's so interesting about the time right now? Well, we have these sort of edicts that a lot of this is moving to GSA. So on one hand I get incredibly nervous, no offense, but I get incredibly nervous when we say, let's put all our eggs in one basket and trust that this organization is going to be able to meet the discrete needs of all of these other [00:31:00] organizations.
Jonathan Mostowki: On the other hand, if we're looking for common solutions to common problems, the logic is there, right? You know, if we wanna say, Hey look, if you wanna buy a ai. This is the organization, organization that buys ai, right? They have the standards, they have the process, they've done the research, and when they give it to you, you can trust that it's usable.
Jonathan Mostowki: That gives a little bit more comfort, I think, than. This bespoke group cobbled together an AI solution with some existing capabilities, some custom solutions, and so on and so forth. And we're the only ones who are the first ones using it. So I think there, the, the, the transition of technology buys or it buys to GSA is an immensely.
Jonathan Mostowki: Important but risky opportunity for the government and needs to be taken with a level of seriousness and caution that it deserves. But if done correctly, no pressure. It, it, it can [00:32:00] really solve a lot of the things we're talking about of.
Kaprice Tucker: I'll just say, um, I think what you praise you'll promote. So, you know, if you find any little example of somebody taking a little risk or thinking outside the box or trying a new feature and you praise it, like it was the greatest thing, whether it's succeeded or not, whether it Yes.
Kaprice Tucker: And that it's hard to do. I'm not saying that we had that figured out. Yeah. But it's, but that's important. Whether it succeeded or not. That's right. So that you permeate the ability to take those risks and to try new things and to actually change the language in a class. Yeah.
Bonnie Evangelista: And it can, uh, I was just ta I've been talking about this last two days with different people.
Bonnie Evangelista: Tim, this morning we were talking about this. Um, my one wish was if, if somebody tried something, I said praising the failure. And maybe that's not the right language, but. Praising the fact that you learned something. We tried something and we learned and we're gonna pivot or [00:33:00] something instead. But that's the part people fear retribution from is if it's not gonna work, then they have the target on their back 'cause they did the thing that the system says you shouldn't do or you know, tells you not to do, et cetera.
Joanie Newhart: Yeah. So, um, we're gonna ask one more question of these guys and then we're gonna let Sariah ask some audience questions. So if you have any questions, get 'em, get 'em in the app. Um, so I loved Bonnie came up with this during our prep. What is your prediction? Like? I'm on the edge of my seat here. What's gonna happen?
Joanie Newhart: Um,
Bonnie Evangelista: kick us off. Okay. I'll start. Uh, 'cause I've thought about this for a long time, especially in acquisition and procurement. I don't think paper proposals are gonna be a thing soon. You, you alluded to this, you know, we've got machines writing proposals, we've got machines eventually, hopefully evaluating proposals.
Bonnie Evangelista: I think it's going to. Create less relevance to paper proposals and we're gonna have to create more environments where you show us something. So for [00:34:00] services that might look like oral presentations, but I think for technology, there's gonna be real no kidding environments that you have to play in to demonstrate something meaningful before you move to the next phase or, or something like that.
Bonnie Evangelista: So I, I think that's a real thing. I also think we're, um. I lost my train of thought. I was so excited about no paper proposals. I just lost my train of thought. I'll, I'll go, I'll, I'll defer to my, to my colleagues over here.
Jonathan Mostowki: Uh, so yeah, a hundred percent on that. I, I think a, a few things. One. This concept of buying AI isn't even gonna be a conversation.
Jonathan Mostowki: It'll be like, we're buying software code. That's just what it, it's going to be, right? Uh, low code, no code, or just, those are gonna be like old concepts. It's, it's just what technology's gonna be. Uh, not to continue to be a naysayer, but I, I also think some companies are gonna win. [00:35:00] And like, there's not gonna be a million like AI acquisition solutions.
Jonathan Mostowki: There's just like, there's Compu Search Prism and like D two P or whatever it is. Uh, you know, there there's only going to be a couple of solutions that the entire government is going to adopt when it comes to the acquisition process. Um, I agree. The whole concept of like receiving a written document is not gonna be worth the digital paper.
Jonathan Mostowki: It's printed on. Right. People don't even have printers anymore. Did you realize that? I just learned that. Um, so, you know, it's, that's not gonna be the way to evaluate, so it's really going to come down to, and I love this like, user experience of everything, right? It's because the technology piece is gonna be solved.
Jonathan Mostowki: Like, don't let your kids grow up to be software engineers, unfortunately, right? Like computers are gonna write software. So it's going to be how do you make your computer written software. More valuable, like the whole, this came up again and, and this was a really good keynote discussion guys. I was, you guys touched on everything [00:36:00] like, like how are you going to evaluate based on our current models of software engineering levels of effort?
Jonathan Mostowki: Like I said, I, I have a client who, the company is 10 people and they are delivering software on multiple contracts. 'cause they use AI to write the preponderance of their code. So, you know, these old models of 500 software engineers to. Solve a problem for maybe CMS for example, you know, that that's, that whole model is broken.
Jonathan Mostowki: And so then we have to think the companies that are built on those models, I mean, they'll adapt. We know they will adapt. They have the money, they have the wherewithal. They have the, the footprint to adapt. So what are they going to look like? And I think that's going to do a lot to shape the future of, of these deliveries.
Kaprice Tucker: How about you Caprice? I, you know, I. Play around with this in my mind all the time, and it's, I'll probably date myself, but we're talking, you know, Flintstones to Jetsons, so to speak. Uh, and you know, [00:37:00] probably the biggest thing for me is the no paper. You know, back in the day it was, oh my goodness, you know, we had storage of all these paper contract files and now, um, electronic documentation is.
Kaprice Tucker: Everywhere all the time, you know, and I, I do believe that it's really no paper, um, when it comes to proposals and evaluations and, you know, you have your electronic documentation across the board. Um, but I also would like to see some kind of way where we are removing the, I'll say, um. Potential conflicts of interest.
Kaprice Tucker: You know, I was talking to, um, our business integration director and saying, you know, one of the things that we talk about is, you know, if you have somebody that's building the AI. And then they're competing on the ai. How do you deal with that? And I think, you know, that too will be a [00:38:00] thing of the past where you work through any kind of potential conflicts, where it's just the run of the mill everyday thing.
Kaprice Tucker: And we know how to deal with all scenarios.
Joanie Newhart: Don't take my paper, please don't take Ani circling back to you. Why?
Bonnie Evangelista: Uh, I'll end with, um, I think we're gonna see more. I'm gonna say marketplace type models. Not necessarily the methodology I was referring to earlier, but like it's, I, I, I look at open source type communities that it's all based on experience and or like how to add value to the end user.
Bonnie Evangelista: And then there's a whole community around like just fixing problems around that. And I think I am encouraged, I saw somebody actually post recently, there's that marketplace model I was talking about. Although it's a methodology, there is a technology side to it because like someone has to, there's an intake process for like receiving all the submissions to go into the [00:39:00] marketplace, and then there is an evaluation process that's aided by technology.
Bonnie Evangelista: They're not doing the evaluations, but like they're recording the evaluations and whatnot. Somebody open source the entire technical side and was, and was challenging the community. Make it better, somebody just make it better. Then we'll figure out how to sell it back to the government. And I think there's gonna be more models like that in the future where like, rather than buying piece and parcel solutions, like we're just gonna have these marketplaces of, and I think the architecture is really important to get this right.
Bonnie Evangelista: Um, where if and the data access, um, too, um, but how can you create communities of. Eng, I don't know if they're engineers of, of people who are building this software and they're, and it's kind of being done without the government dictating what problem to solve or what requirement to solve and which is kind of how it's done in the commercial space.
Bonnie Evangelista: So I think that I, I am encouraged that, that, that seems far reaching, but I think that will happen.
Joanie Newhart: So just during this panel, I've gone on a journey from fear to optimism, [00:40:00] so I'm feeling better now. Thank you. But we'll turn it over to Sariah. Take us to the finish line. Thank
you, Joanie. Thank you so much.
I did wanna make a couple of comments. One, I think leadership plays a critical role. You will not. People will not do these things. They are gonna have fear if they don't have that top cover. I really emphasize that. You gotta live by that and it's getting the right leaders in place. And then the other comment that I was gonna make, acquisition is a team sport.
Yeah. So the fact that other people aren't going out there and innovating, you know, for example, like when I created the pill, we did get some people that actually started bringing us ideas from outside of the organization because it is a team sport. Yeah. But they have to know that they have top cover in government.
One of the questions that you're gonna get right now is, the biggest thing that contributes to the fear is that we live in a risk averse culture. Why not because of the people in this room. Not because of any of you. Not because of me when I was in government, but because we have all these oversight organizations, because we have a [00:41:00] congress who reacts when something bad happens and they legislate a solution.
That's what we'll call it a solution, right? So the question is, agency policy and compliance contribute to slow adoption of AI and the fear factor. What are some effective approaches to reducing the policy barriers to help accelerate AI adoption? Told you, I was gonna throw you a tough question.
Kaprice Tucker: Uh, I guess I'll start with using AI to help with policy.
Kaprice Tucker: Uh, you know, um, I mean, it's funny, but at the same time it's true. You know, yes. You're using AI to whether it is, um, understand, uh, if you just think of all the different executive orders that we've. Recently received, you know, breaking those down and, and not feeling like that you're, you know, on a wheel, um, to understand what you have to do.
Kaprice Tucker: But using AI to break down the requirements and then also to [00:42:00] reduce the amount of regulation and policy and still be effective. And that is a change in culture, but it's not so far fetched that, that we can't get there. Yeah, definitely. Anybody
else?
Jonathan Mostowki: Yeah, I mean. I, I, I think we just experienced a little bit of, we got a whole bunch of AI policy and executive orders, and then we had a whole bunch of AI executive orders removed.
Jonathan Mostowki: And, uh, you know, I mean, I, I, I think the government has to have a, a single perspective. Of what it will accept. Mm-hmm. Not overly specified. 'cause look, we're, we're talking about a lot of the pros of ai. There are risks. Absolutely. Okay. So this is not just, you know, this is just fun and cued and let's see what happens.
Jonathan Mostowki: Like this stuff can get screwed up. Mm-hmm. So there, there does need to be a standard of adoption, but what there doesn't need to be is 500 standards of adoption. Agreed. You know, like I think the SER program is a really good example of this, right? The SER [00:43:00] program is a thing created by Congress outta statute to make it easy for small businesses to do business with the government.
Jonathan Mostowki: Right? Right. So you have the the SER top level set of rules. Mm-hmm. And then let's just take the Department of Defense. You have the DOD set of SER rules. Mm-hmm. And then you have the Air force set of rules. Yep. And then within the Air Force you have their csso Sbra pipeline, and then you have their other SER pipeline.
Jonathan Mostowki: And I'm not knocking them done a lot to make it good. What I'm saying is, if I'm a small business responding, I have to know all of that. Right. AI can help, but I have to know all of that. Right. And so I think we need to do the same thing with policy. Okay. Just be clear, keep it simple, stupid. And, and, and stop there.
Bonnie Evangelista: So career. So I don't wanna be terribly, I hope this doesn't sound too reductionist, but along that same thread, like what you were just describing is actually a little bit of a myth. There is no air force. Uh, DOSD or whatever policy of ser, but what's happening is everyone's interpretation of the SBA SER [00:44:00] policy is different in each one of those organizations.
Jonathan Mostowki: They all have their own pride, and so
Bonnie Evangelista: asking the why is incredibly important in this culture or in this era that we're in. Why does the policy say that? Like who says we can't do it that way? Or why are we still doing it this way? And if a policy comes up like we need what Jen Palka talks about in her book Recoding America, that traceability back to statutory intent.
Bonnie Evangelista: Is the policy that's implementing the statute actually like getting the outcome Congress wanted? Or are we just like overthinking the problem like you're suggesting and like just making this harder? Because if you. If you just look at like sr's a great example. OTAs are another great example. There is a statute and a policy period, and then there might be some local guidance and procedures.
Bonnie Evangelista: But what's happening is 'cause there's a lack of training and all these other things that we're talking about, culture, everyone's like not sure and afraid. So they come up with their own rules on how to handle [00:45:00] it and it may not actually jive with the outcome Congress was trying to get. So like I would offer like.
Bonnie Evangelista: Policies. Not a lot. It's a suggestion. It's a, and we should challenge our policies if they're not getting us the outcomes that we want or need to have right now. So that, that's my reduction as you Yeah. I agree with
you. I, I, it's not the policy, it's how people interpret the policy. That's really what's killing us out here.
Um. So I am trying to find a way to word this question a little more positively. Um, well, 'cause Josh, you kind of hinted to this, um, what are, what are some of the concerns or how can we use ai? Let me ask this, this way. How can we use AI to train or retrain the workforce to enable them to use ai? Can, how can we use ai.
To train and enable that workforce to use AI to remove that fear? Well, this is a real [00:46:00] philosophical question. I mean, yeah, I mean, I don't know that,
Jonathan Mostowki: I don't know that I, my answer's gonna actually answer that question because I don't know that you use AI to train people, but what you do is you expose people to tolerable levels of risk to make them comfortable.
Jonathan Mostowki: It's like behavioral therapy, right? Just exposure. Yeah. Exposure. Yeah, conversation risk removal, you know, and, and, and I think you have to move at a pace. So like, I always tell my customers, you gotta meet your customers where they are. Mm-hmm. Right? So you may have a very bold, innovative leaders, like, we're going to do this.
Jonathan Mostowki: Right. And you're gonna have early adopters all the way to, to laggers. And, and within that parallel, you gotta figure out, you know, the common denominator. Like where are you gonna snap the line and say, this is how we do business. Mm-hmm. Either. Get on board or get off boat, you know? Uh, and then you gotta have some tolerance for, you know, some people are gonna do it, you know, a little bit differently and some people are gonna fully embrace it.
Jonathan Mostowki: Sure.
Bonnie Evangelista: I'll, I'll give an example. Somebody [00:47:00] recently shared with me, they talked to chat GBT on their commute to work. And I was like, wait, what? You can do that? I didn't even know you could do that. And so I figured, so that one little suggestion kind of expired me. 'cause I was like, I, I am very much like a.
Bonnie Evangelista: Avid podcast listener for example, like that's my method of learning really well is in conversation and, uh, figured out how to use it in the app. And, and I have full on conversations about like the most random topics. Mm-hmm. Not just acquisitions. So like, I do think there are ways like that, that's a very small mm-hmm.
Bonnie Evangelista: Thing that you can do today where you can use AI to learn. About either your craft or your trade or just something totally random. And there's probably more to it that I'm not even thinking of, but that that is another example of exposure, though. That's a very low hanging fruit way to expose somebody to the tool.
Bonnie Evangelista: And it doesn't even have to. Another way is, um, a, a way to help with the exposure piece is take away the job function. Mm-hmm. And just like, [00:48:00] uh, the more Alexis Bonnell, who now works at OpenAI, she used to work at Google and she was the A-F-R-L-C-I-O. I learned a lot from her. And she's at, when she was at Google, she found that the more you could find relatable ways.
Bonnie Evangelista: Um, to use cases for people to experience that helped with any kind of technology adoption. So, for example, whether it's chat, GBT, giving them scenarios where they can use, like, like come up with a, a workout plan or mm-hmm. Uh, find lowest airfare for your next travel or something. Make it like, not even related to work.
Bonnie Evangelista: Right. And just let them have that experience and then they, they go from what she calls the ho hum. Like that fear that like, I don't trust this or whatever to the aha moment. Yeah. Yeah. I would say that I
Jonathan Mostowki: think you could, sorry just for, I apologize. Uh, you could have adaptive learning, so where the, you only have to spend time learning the stuff you don't know and AI can help [00:49:00] direct that.
Jonathan Mostowki: That would be the only,
Kaprice Tucker: I was just gonna throw in, um, you know, we had a recent CO just volunteer to use our DOI. Chat, GPT on different use cases in the acquisition process and just opening people's eyes on, oh, I can use this to understand, you know, the rule of two or, you know, whatever it is in specific acquisition areas to break down things.
Kaprice Tucker: And that was just part of learning where they're like, oh, I didn't know I could do that. Just more exposure. Yeah.
No, thank you very much. I know we're out of time, by the way. You won the, uh, question. Pool here. There's like probably about 30 questions in here. Um, so I'm sure that you're gonna get grabbed on the way out the door, but I, I do wanna thank this panel Caprice, Josh, Bonnie and Joni.
I've known these individuals for quite some time and I value and appreciate their opinion, so please join me in thanking them for their [00:50:00] insights.