AI Leadership Insights: AI’s Role in Streamlining Government Services with GovSignals’s Derek Hoyt

Hello and welcome to the latest edition of the Techstrong AI video series. We're here with Derek Hoyt, who is CEO of Gov signals, and we're talking about how AI might be applied to, well among other things, make contracting for the government a lot more efficient, and that has implications for maybe how we're gonna save a lot of money on government spending in the future. Hey Derek, welcome to the show.

Uh, thanks Mike. Appreciate problem. So what is the possibilities for using AI in terms of, uh, automating government contracting?

The whole process seems rather labor intensive still and costly. Um, but is there an opportunity here or is it more wishful thinking? What do you think?

We strongly believe aircraft signals that the AI is able to unlock a lot of opportunities. So you're right in that a lot of processes right now are very manual, so very, but also very linear in their their steps. And so what AI is unlocking is the ability to, um, streamline a lot of these sequential steps and collapse that time it takes to execute these repetitive tasks.

Um, that's very much where signals, right? It comes into play is because we're redefining who's winning these trillions of government contracts each year with our first AI platform that automates the school for process of finding and bidding on government funding. What would it take to set something like that up?

Because we're some folks in the new administration are making some very aggressive promises about how much money we can save, and, um, is there a gap and, and what is the effort required? So from a implementation standpoint, so like for example, you know, we support, um, both government contractors and, you know, and government agencies themselves. The system has that ate.

So let's start with a, a classic, so like a government contractor side so they can get up and running and using the system. We have folks literally gleaning in sites and using it within, um, a week and being able to, we aren seeing tangible impacts. The other side, the flip point is then we have those larger enterprises like government agencies or larger corporate enterprise, uh, clients like that.

We have, this is where, um, it takes maybe a little bit, uh, a little bit more time upfront. It's more of getting the data a plug into the system. But once the data is connected, that's where you really start to see the value of these workflows being optimized by the AI systems, um, really taking effect.

Wouldn't we therefore reduce the number of employees that we need in the government or would they be kind of reallocated to maybe tasks that are, um, higher value? What would be the impact? Well, like everything, there's a lot of flexibility.

So for instance, um, when you look at it, I'll use a tactical example, like for like corporate, like some of our corporate clients. So larger enterprises, they, when they have our access to our system, they're actually able to repurpose, uh, a lot of their teams in order to be more effective. So they're saving more time.

So that actually increases their output so they don't have to hire more staff. They are able to glean more efficiencies with the current step that they have to maximize their output. What's your sense of what are the hurdles to making this happen?

'cause it seems like there are both technical and cultural issues that would be at play here. Well, like anytime you have change, right? So change management is always, um, there's always and maybe potentially initial friction.

And so one of the things that we always recommend is like a crawl, walk, run approach. So very much start with, um, a smaller cohort to demonstrate and learn the nuances, and then you begin to expand, uh, quickly. Um, once you have capitalized on tactical use cases or demonstrate like, yes, these type of processes work, um, versus, um, a large, a large push out, uh, or new systems and capabilities, um, increases the friction of adoption, uh, because it's based on a top down mandate versus a, um, uh, like per user adoption and acceptance of the value that these, uh, that a system like up signals provides.

So given all that, how big a journey are we looking at? I mean, it seems to me that this might be something that takes multiple years, there'll be benefits along the way, but it's not like we're gonna flip a switch and have AI drive everything overnight. No, I don't think the, an overnight, um, transition is what, um, I think anybody would expect in the industry.

I think what we would see is as these systems continue to evolve and expand in their capabilities, um, and also the adoption, um, and integration with human-based, um, processes, um, I, I see a slow but methodical transformation, um, at scale. But what's exciting is that we see very rapid add, like at a per client basis, we see the rapid tactical adoption and acceleration, um, of like traditional workflows. So for instance, finding bidding opportunities or scanning proposals or writing proposals.

At these tactical levels, clients are getting rapid responses and results within, like I was mentioning, like within like a week, uh, onboarding to these type of platforms. So I think what we see is that as more and more of these use cases of success start to permeate in the various, uh, industry, both in the government contracting side and the government side, I think that's when you'll see that o that overall adoption of these new capabilities. Are there particular AI models that are more, um, in tune with this particular task?

I mean, or are they gonna be a mix of these things going forward and we're gonna have all these various AI agents that kind of connect to each other somehow? Yeah, from gov signal's perspective, uh, uh, we use an ensemble model. So we're not just one specific model, uh, based, we don't use just one specific model.

We use a collection of various models, various agents, um, knowledge graphs and a lot of different technology in order to really glean the value. Um, just throwing data at a model is not enough. That's not gonna work.

Which you need is a platform that acts as like a, a harness basically that is able to interact and understand the nuances of these specific human-based workflows that are being done now manually, and then be able to translate that into a, um, clear set of instructions for the models to and agents to work with. Being able to just throw something at a model and then hope that it, it comes back with a clear response. Like a, for example, like a chat GPT based interaction, just sending something and expecting result back, you're not going to get to that depth that's required to actually drive business operations.

Like a Chachi PT type equivalence, like a one shot interaction pattern might be good for summarization or like high level characterization of activities, but to really drive like we need to write a highly compliant proposal that's based upon all these various requirements, um, that also needs to be nested to some of our limitations and resource allocation that requires an ensemble of models and agents working together inside a platform to deliver results Soon. Does somebody have to write the orchestration framework to allow those agents to kinda, uh, collaboratively create a process and also something that we're gonna insert humans in the middle of it when necessary? Yes, 100%.

And that's exactly why we found the gov signals, uh, to be that orchestration layer to establish the connectivity best of breed models in the industry with the best technical processes for our clients and deliver that value every day. So how will this work out where if I have a set of agents that are optimized for a specific workflow and that workflow is diametrically opposed to your workflow because I'm a buyer and you seller or a con contractor or whatever it may be, how will those things come together and kinda do something positive or will they simply cancel each other out? Actually, I can give you, lemme give you a tactical example.

So some of the things that we're starting to see in the government space. So we're able to use AI assisted based workflows to write compliant proposals very quickly now using company's information to tailor based upon the requirements. Now when you have that output, that that document, that is the output of that process, the receiving entity in this case, the government agencies are now receiving this document and they can now use AI systems to then scan that document in order to determine, um, the eva, did it meet the evaluation criteria, give relative scoring, et cetera, et cetera, and be able to do that at scale.

So the output of one system then becomes the input of another evaluation based system. So it doesn't necessarily mean that these have to be diametrically opposed, it's just different functions that the AI agents used based upon, um, in this basically two-sided marketing. Ultimately, how interactive is all this gonna be for the average person?

Because I think one of the issues we have today is that people feel fairly removed from their government services. I gotta go stand in line, take care of a lot of issues and almost dedicate half a day anytime I engage. Is that whole experience gonna change?

I think fundamentally, yes. As we continue to see AI systems become more adopted, especially for the delivery of government services, uh, the natural byproduct is going to be an increase in efficiency. But the other piece to consider is that there's gonna be changes in process.

So for instance, like working with our larger clients at Go Signals, uh, they, because of the time savings that they receive that they're seeing and collapsing the effort that used to take maybe two weeks, they're now getting done in a couple hours. This is opening up more capacity to provide either additional quality to the services or opens up ring for offering additional value somewhere else inside the organization. So this is unlocking, uh, AI is unlocking a lot of, um, potential, uh, worth and value inside the existing services that just wasn't there before because that time was capped on manual efforts.

So I think we're gonna see some really fundamental changes in government services as we just unlock a lot of this manual work. So what's your best advice to folks who are running these agencies today about how to get ready for all of this? Wouldn't you wish that they all knew now versus having to wait for you to tell 'em?

I think the big thing is, um, continue to work with, uh, clients that understands, uh, how to do like FedRAMP, uh, moderate or high certifications, understand the process for security and data handling and working with clients that understand how to deploy solutions at scale. So if you working with, um, companies that understand both of those, both the security of data and the criticality thereof and then how to deploy at scale enterprise based solutions, that's gonna be where you really get the value at across an agency versus just prototype examples, uh, throughout like different one-off projects. Um, if you really want to see an at agency scale level type change, All guys you're heard in here, AI is coming in government generally for the better, but we may be some hiccups along the way, but hey can't be any worse than it currently is in some regards.

Right. Derek, thanks for being on the show. Thanks Mike for having me.

Thank you all for watching the latest episode of the Text Drawing that AI series. You'll find this episode in those on our website. We invite you to check them all out.

Until then, we'll see you next time.