Gemini for Government - Six Five On the Road

How are federal agencies leveraging next-generation AI while balancing stringent security and modernization requirements?

‍

Host Patrick Moorhead is joined by Google Public Sector's Chris Hein, Field Chief Technology Officer, for a conversation on Gemini for Government. The discussion addresses how AI platforms are enabling federal agencies to transform mission-critical workflows without compromising on control, flexibility, or security.

‍

Key Takeaways Include:

🔹Comprehensive AI Capabilities: Gemini for Government offers enterprise search, video and image generation capabilities, and AI agent functionality in a single platform, addressing diverse agency needs more cohesively than siloed AI tools.

🔹Security and Innovation via Commercial Cloud: Leveraging Google’s accredited commercial cloud allows agencies to access constant innovation while adhering to federal security standards, avoiding the delays often seen in dedicated government clouds.

🔹Avoiding Vendor Lock-in and Customization: Agencies can adopt, customize, or build their own AI models and agents on Gemini, supporting mission-specific solutions and reducing dependency on proprietary platforms.

🔹Productivity and Collaboration with Security: FedRAMP High assistants for Workspace and Agentspace bolster productivity, secure knowledge access, and collaboration for federal workers in daily operations.

Learn more at Google.

‍

Watch the full video at sixfivemedia.com, and be sure to subscribe to our YouTube channel, so you never miss an episode.

‍

#GooglePSSummit #GeminiforGovernment 

‍

Or listen to the audio here:

‍

Disclaimer: Six Five On the Road is for information and entertainment purposes only. Over the course of this webcast, we may talk about companies that are publicly traded, and we may even reference that fact and their equity share price, but please do not take anything that we say as a recommendation about what you should do with your investment dollars. We are not investment advisors, and we ask that you do not treat us as such.

‍

Patrick Moorhead: The Six Five is On The Road here in Washington D.C. at Google's Public Sector Summit. The talk is AI. We certainly know all of you out there that from a consumer basis and a business basis, what needs to happen now when you get into government, types of workloads and transformations, security, resiliency and governance are even more important than these other markets. Google is really taking the market by storm here. I've been tracking them since the inception of Google Cloud and they've really been making inroads here. And one of the key technologies here is Gemini. And Gemini is the portfolio name for a set of technologies that I got a great demo of today. And quite frankly, I even created my own agent. Even me. Can you believe that? I can't. But to talk about this is Chris Hein, CTO from Google Public Sector. Chris, welcome to the show.

‍

Chris Hein: I'm delighted to be here.

‍

Patrick Moorhead: Yeah, it's been great conversations. Saw your keynote that you were part of and I was really struck by. I'm at this very simple interface. Let's pull down what data that I want to use and let's create an agent again. Like I said in the lead up, even somebody like me could do this. So I want to just dive in. Gemini for Government, how is it helping the public sector to fulfill their missions? It's a broad question, I know, but I want to start this conversation broad, sure.

‍

Chris Hein: So what Gemini for Government is intended to offer is we've created it to really be a front door to the rest of all of our AI services. And so it becomes kind of this overarching way of thinking that we're going to build a user interface that makes it so that every single worker can come in. It needs to be easy. Right. Folks like yourselves need to be able to make an agent, but it really is intended to give access to all of the best of what Google is creating when it comes to our artificial intelligence products. And so whether it's our Gemini models, all the way to our image generation and video generation models, all of that just lives in this one spot that makes it easier for folks to be productive there. But the way that it starts to really kind of accelerate things, especially for government agencies, is that it doesn't stop at just being kind of a personal chat bot that might just make you a little bit more effective. What we're trying to do with it is we're trying to make sure that a. We're connecting, as you pointed out, we're connecting into the data sources.

‍

Patrick Moorhead: Yes.

‍

Chris Hein: Where is your work done today? Right. The government does not change quickly. And so there are all sorts of data silos that exist and they serve their purpose. What Gemini for Government allows us to do is to connect to those, make those part of your overall experience of AI. We're going to bring AI to your data, then the next part of it is the agentic workflows. And so this is where we start to move beyond just personal productivity and talk about enterprise productivity. Because the way that AI can actually start to change how you do work is when you start to think about workflows themselves. And how do you implement an agent that can take on like a discrete task? Right. There's gotta be something that it needs to do, a goal that it needs to accomplish, and, and the data sources that it needs to go do that. And so we've got a variety of ways that you can then inject these agents into this Gemini for Government interface so that whether it's you building your own, you know, just using like, you give it a goal, you ask it to build a prompt for you, you tell it what data sources it needs and now that's there, you've got an agent all the way to your enterprise, creating very custom ones that are gonna have direct access into specific tools and orchestration events that can really take you from just a, again, personal productivity to really transforming how a government agency can work.

‍

Patrick Moorhead: Yeah, I was really impressed that you even showed data sources from incumbent cloud vendors out there. I think that's a great confidence booster for people and obviously it's valuable because people have heterogeneous architectures that they've been leaning on for decades. Talk to me. Your solution seems very integrated. Can you talk to me about the end user benefits of aggregated versus, let's say disaggregated AI tools?

‍

Chris Hein: Yeah, and I think what we realized at Google Cloud was that it was going to be absolutely necessary to your point of saying this is going to be a multi cloud solution. You're going to have a hybrid environment with a whole bunch of other SaaS platforms that you're doing for your day job. You're going to have, you know, maybe you're a Microsoft Office user and we want to make sure that you still can get the best of what Google's AI can offer, regardless of what your productivity platform looks like. And so what it starts to look like. And this is really important because what we're finding is that if you just replace six different SaaS applications with six different chatbots that I log into each one and each one doesn't know about each other and they don't have the same common data source or context for me. And then it just becomes sprawl.

‍

Patrick Moorhead: We call that swivel sharing between multiple. Like okay, this system, I'm going to go over here, I'm going to go over here and absolutely is a time sucker and I think bad for morale.

Chris Hein: Absolutely. You see those folks that have six different monitors up and each monitor has a different thing that they're responsible for. And so how do you start to lower the amount of cognitive work somebody has to do being like, oh wait, I go to this system when I need to answer this question. Right. But instead of starting to say, okay, no, I've got one interface that's connecting into all of those systems and gives me the ability to kind of do work in one cohesive way versus very disparate and kind of all over the place.

‍

Patrick Moorhead: There's been a lot of talk here about your accredited cloud and can you just use very simple terms for people. What's the benefit of it compared to the gov cloud that we've been using for decades?

‍

Chris Hein: So the biggest difference for Google we came to this conclusion that gov clouds have very specific limitations. One is resiliency. So you're only going to be in the very clusters that are already government specific. If there is an outage or if something happens to that specific cluster, there's no backup. Right. There's just no way for you to balance that out into the rest of the ecosystem. Whereas the commercial clouds you can be across to the entire country. Right. And you can make sure that regardless of if you've got a hurricane or any event that's affecting one region, you've got that level of resiliency that doesn't exist in other places. But then the other big problem that we've seen is that your tool sets fall behind really, really quickly because it's not necessarily in the cloud vendor's best interest to keep that up to par with everything else that's available for commercial use. And so from what Google has created, we've decided we're going to take our commercial cloud and we're going to accredit it at the FedRAMP high status. So what that has meant is that it takes every single commercial offering that we've got on the truck and it makes it so that government customers don't have to say, oh, I'm getting the second rate version of all of these same services, they're getting the same top rate. And then in today's AI world it matters even more because you know, maybe you need access to Nvidia chips, right, to be able to run a specific AI workload. Well, great. I've got them across my entire commercial cloud. I'm not capacity constrained on that. Whereas in the gov clouds, they're extraordinarily capacity constrained on things like GPUs.

‍

Patrick Moorhead: No, this makes sense. I think one of the advantages that you have is the full stack and then out of the other side of the mouth that's, hey, but we're flexible. You want to bring your own model, you want to bring your own agent. That's really good. Good for you. Can you talk about it? I mean, I think sometimes I ask questions, I know the answer. I'm an analyst, I need to know these things. But I really want to hear in your words, what's the benefit of having both of those modalities?

‍

Chris Hein: So a lot of it is that we don't think of the kind of AI transformation as being one model is going to rule them all. This is very much going to be a multimodal world for the foreseeable future. And so there are times where we're obviously very proud of our Gemini models.

‍

Patrick Moorhead: Well, and quite frankly, you're crushing it on at least the benchmarks that are out there. And I think your API calls are the fastest slope of a relevant model out there. So congratulations.

‍

Chris Hein: Oh, it has been, it's been a whirlwind, to say the least, the last couple of years. But what it ends up becoming is, you know, we've got best in breed models, as you point out. Like we've got best in breed models, but even inside of a best in breed, there's still really a lot of room for variance. Right. And so there's some models that are out there in the ecosystem that are fantastic coding models. Right. People love coding with the CLAUDE models. And what's amazing about our viewpoint on this is that you can go ahead and access those CLAUDE models through our cloud platform. It's right there. And we get the moment that Anthropic releases a new version, they're right there and available for that customer to turn on and use. And so each rendition of these models, as they continue to accelerate, there's always going to be a reason that you might choose one or another, and we don't want to make it so that you're kind of feeling this lock in. Right. Like so much of what software has been over the past 20 years has been, okay, make your smart choice, because that's what you will be experiencing for the next 10 years of your life. And we really want to get to a point where, look, we can. We can operate against all of these different models and modalities and really give you a lot more choice.

‍

Patrick Moorhead: Yeah. And the way that I'm assuming the way you keep everything secure is you are managing the models and the agent through. Through a centralized orchestrator.

‍

Chris Hein: Yes. And a lot of that becomes important to say where the data is, is gonna be where things are most important. Right. And so you're able to kind of know that you have this secure foundation that you're working off of. So we talked about accreditation, and that's an important component of it, is making sure that we can validate to our government customers that we are keeping that data secure. So then every API endpoint that has access to that is, is going through the same zero trust environment that Google has assured to the government. So that you have that flexibility to change from one thing to another without worrying about going through a whole new security audit. Because it's already been done in that first one.

Patrick Moorhead: It almost seems too good to be true. Which, by the way, that's a good thing.

‍

Chris Hein: Yeah.

‍

Patrick Moorhead: That people will poke on that. I hear a lot of other companies claiming that they have the same thing. Good answer.

‍

Chris Hein: I've heard those same claims.

‍

Patrick Moorhead: Okay.

‍

Chris Hein: All right. I'm suspicious.

‍

Patrick Moorhead: Okay.

‍

Chris Hein: You know, the thing that I don't think anyone can claim that you see out of what Google has brought to this is that, you know, we talk a lot about our full stack. Yeah. And this is where we really have differentiation that it's going to be hard for anybody else to make a viable claim across. Which is to say we have been building because of Google's own needs, the world's best AI infrastructure for over a decade.

‍

Patrick Moorhead: It's planet scale.

‍

Chris Hein: Absolutely. And as you look at that, We've got seven or eight applications that are used by more than 2 billion people each. That requires a level of sophistication at that infrastructure layer that our public sector customers get to take advantage of. And as you move that up into the Google DeepMind research teams, that is a level of research sophistication, a number of publications. You know, we invented the transformer model here at Google. Right?

‍

Patrick Moorhead: Yes. I was reminded. Reminded my audience of the last two videos I shot here.

‍

Chris Hein: So good. Yes. So we're all the way at that. And then just creating these agentic solutions kind of lives on top of this stack that we already have this expertise across that entire thing. And so while there's other vendors that are doing good work. Right. Like there's interesting things that they're doing. It's very difficult to find anyone out there that can give that breadth of the overall portfolio.

‍

Patrick Moorhead: So I want to end on a high note, a FedRamp high note.

‍

Chris Hein: That was a very good point.

‍

Patrick Moorhead: See what I did there?

‍

Chris Hein: Yes.

‍

Patrick Moorhead: Now, seriously, Fedramp High assistance in workspace and agent space. Can you talk us through that? What are the differential benefits that you're bringing to your customers?

‍

Chris Hein: Right.

‍

Patrick Moorhead: Related to productivity, security and even knowledge access?

‍

Chris Hein: Yeah. So the big thing there. So the FedRAMP high accreditation gives us kind of that authority to go operate. Right. And it's in a way to go and really present to agencies, you know, if you look at workspace, this is the first time that they can really look across the entire vendor market from a productivity tool set and say that they have legitimate choice, you know, in what they're going to have their agency workforce use. And. And so by adding that to their kind of portfolio of options, a you reduce vendor lock in which that's always a good thing. But also the security of our system is paramount. And so we've seen, you know, there have been breaches in many, many government systems over the past couple of years. And so by giving them this incredibly secure offering that is going to give them the best security possible on that kind of Fedramp accreditation, so that they have that ability to go per purchase and work with it. That's huge from that perspective. And you flip over into our Gemini for government being also at that Fedramp high level. What's important about that is that, you know, there's lots of AI vendors right now and it's very difficult to go through these accreditation processes. And it's the fact that we have been putting in the blood, sweat and tears, sometimes literally, that is giving us this kind of ability to offer to them to say, yes, we can give you this and it's available right now. This isn't some kind of. Maybe in the future we'll get to that compliance perspective. No, no, no, we're there. You have the ability to start to integrate these tool sets today. And that's a huge benefit for those customers.

‍

Patrick Moorhead: Well, Chris, I want to thank you for the time here. Understanding, better understanding Gemini for the public sector, and quite frankly, helping all of governments to fulfill their missions. By the way, that's why I love verbiage. Missions are awesome. And that's not just a military mission. Every one of your customers has a mission, whether it's educating, protecting people and everything in the middle so I'm going to thank you for your time.

‍

Chris Hein: Thank you so much for having me.

‍

Patrick Moorhead: Thank you. This is Patrick Moorhead signing out for Six Five. We’re here on the road in Washington, D.C. at Google's Public Sector Summit. It's been a great show so far. Check out all of the summit content and also we have a lot of other Google content as well. Hit that subscribe button. Thank you so much.

‍