Box and the Future of Trusted AI in the Public Sector

In this insightful conversation, Brett Conrad, Vice President of Public Sector at Box, joins Triad Technology Partners to discuss the tangible shift of generative AI in government from theoretical exploration to practical implementation in 2026. Moving beyond the initial hype and questions of 2025, the discussion centers on how agencies are now actively scaling pilots into production use cases. Brett details Box’s strategy, which leverages its deep security and compliance foundation (like FedRAMP authorization) and its role as a consolidator of unstructured data to solve critical challenges. Key themes include the necessity of cleaning up data silos to provide true “context” for AI, the importance of a model-neutral platform, and a hands-on, co-building approach with customers through pre-built AI agents and workshops. The episode provides a realistic and actionable roadmap for public sector leaders looking to de-risk AI projects, drive modernization through “micro-efforts,” and ultimately achieve better mission outcomes.

Relevant Insights

Brett Conrad leads Box’s Public Sector organization supporting over 10,000 US State, Local, K-12 Education, Tribal, US Federal Civilian and DoD customers. Brett is an industry thought leader on enterprise content management, SaaS, GenAI and cybersecurity delivering 100s of digital transformation and modernization projects for customers over his 11+ year tenure at Box. Brett also has deep focus on driving impact in the local community supporting non-profits and NGOs through his work with Box.Org in areas such as youth education, child welfare and homelessness..

What You’ll Learn in This Podcast:

The State of AI in 2026: How the public sector conversation has matured from “What is Gen AI?” to actively scaling pilots and proving out production use cases for mission impact.
The Foundational Challenge: Why consolidating data and eliminating content silos is the critical first step to making AI work, as context is king for accurate results.
The Power of Model Neutrality: Why betting on a single Large Language Model (LLM) is a risk, and how a flexible, multi-model approach (like using Gemini for images, OpenAI for text) leads to better outcomes.
A Blueprint for Pilot Success: Box’s practical, two-step methodology: 1) Building trust through security/compliance education, and 2) Providing the “AI Content Foundation” infrastructure so agencies don’t have to build it themselves.
The Acceleration of Pre-Built Solutions: How Box is packaging repeatable processes—like data extraction, AI-powered search, and security classification—into configurable, pre-built AI agents to speed up time-to-value.
A New Engagement Model: The shift from simply providing technology to “co-building” solutions in hands-on workshops with mixed teams of IT and line-of-business end-users.
Overcoming Adoption Hurdles: Strategies for change management, starting with small, high-value use cases (like automating manual data entry from PDFs) to build confidence and demonstrate quick wins.
The Vision for Scale: How the concept of a “solution library”—sharing proven workflows and agents across agencies—can prevent reinvention and dramatically accelerate public sector innovation.
Box’s Strategic Role: How Box positions itself as an intelligent content management layer and consolidator of point solutions within a broader ecosystem of best-of-breed technologies like Salesforce and ServiceNow.

The Intersection of Governance and Innovation: Box’s AI Future:

“I’ve been in the content management space for a long time, you know, almost 15 years now. I’ve been at Box the last 12 years, and I’ve been fortunate enough to be leading our public sector business for the past 3 years. Public sector was the first fully verticalized industry at Box, and we’ve been placing deep investment into that ecosystem. So, in my current role, I oversee state and local across the United States, K-12, tribal, as well as our federal, civilian, and DOD practices. And we support over 4,000 government customers across the United States. We have a very large footprint in the federal government, where we have nearly a footprint in every executive branch agency. We’ve been placing deep in-region focus to support both state, city, and county customers, and we are really excited about the time we’re in because we’ve been investing very heavily in our platform and are in a great spot to drive a lot of the AI innovation that government agencies are trying to get their arms around and scale and adopt and find success.”

The Shift from AI Questions to Action in 2025:

“It’s been a great year, supporting public sector. I think in the ecosystem, if I look at, kind of, what did we solve this year? So, last year when we were entering, if you went to any government conference, the conversation was, what is Gen AI? What could the impact be for our mission? How do we put guardrails around it? How do we secure it? This year, candidly, we got into a lot of, active conver- conversations, pilots, proof of concepts. We’re now scaling out a lot of production use cases.”

The Challenge of Data Context & Model Neutrality:

“If you want to get the context, you know, what we see from a lot of employees is they’re trying to answer a specific question. Hey, I’m trying to find the associated case files, and I’m trying to run analysis across these case files so I can make a determination faster. But if that file lives in 5 different systems, and you’re not sure what the authoritative source is, you’re gonna have a really hard time making AI work for you… we are a strong believer that there will be no one model for every use case. So model neutrality is a huge, belief and mantra that we base our platform off of.”

Approach to Successful Pilots, Security & Pre-built AI Agents:

“Thank you for hitting on the security and compliance piece. I think we do have a nice structural advantage in the sense of, like, we’ve had a multi-year effort, and super exciting news this year that we hit our FedRAMPI authorization at the federal government… Step two is where we’ve seen a lot of folks struggle in a pilot is underestimating, actually, the core infrastructure, or we call it the AI Content Foundation… And then, an interesting thing that we’ve been doing, and we’ve been scaling this out over the past 6 months, is we started to pre-build AI agents for our customers.”

Hands-on Co-Building & Industry Collaboration:

“On the pilot success, I think what we’ve found to be really successful is we’ve been doing a lot of hands-on workshops, so we’ll come in and, you know, bring, whether that’s bringing laptops in or bringing… enabling a demo environment, we’ll bring in a mix of, line of business, end users, and IT… Now it’s really, like, how do we co-build? How do we refine, based off the inputs that you want?… I’m seeing a lot of, like, we fundamentally, at Box, your point, we’re very focused on delivering the best experience around content management and intelligent content management with the AI capabilities. We’re gonna be a piece of the broader puzzle, right?”

Addressing Concerns & Starting Small with AI:

“I think you’ve seen a lot of public references of, you know, there was a court system where someone drafted a brief, a legal brief using AI, and, you know, that was a big issue. I think everyone is, and rightly so, is still concerned about employees, you know, making determinations using AI without using, without having a human in the loop… I think what we’ve also been focused on is, like, how do we start small? And maybe the areas are, like, we know that employee… government employees spend too much time manually clicking and opening PDFs to pull out, you know, the name, the case number, and then keying that into a system.”

Vision for 2026: Scaling Solutions & Building a Solution Library:

“For me, it’s all about scale, scaling solutions. So we have proof points for many, many government agencies, but I don’t think we make it easy for government agencies to learn about how other folks and similar peers are using us. So, we are very focused on this concept of extending out a solution library of capabilities, and think of it like a… think of GitHub, right? Someone writes some good code, and you want to make it accessible and open source to other folks… now the next step for me is, how do we make it easy, and this is directionally where Box is going, of a solution library.”