Northern Rockies Regional Municipality is modernizing public service delivery and accelerating records requests with Box AI by turning decades of unstructured data into secure, actionable intelligence.
Advances in AI offer profound potential to enhance public service delivery and meet rising constituent expectations. But first, content must be centralized, secure, and usable. The sprawl of government records has long created issues with fragmented ownership, discoverability, access policy, and retention and audit needs. For state and local governments, AI readiness starts with content readiness.
Northern Rockies Regional Municipality in British Columbia uses Box AI to accelerate requests, organize decades of unstructured content, and surface answers from technical documentation in seconds. But initially, Northern Rockies’ IT team banned AI tools across the organization.
This wasn't a rejection of innovation. It was a demand for security, and the first step to a secure, compliant, and measured approach to AI transformation. Here’s their story.
Key takeaways:
- Security and data governance first: Northern Rockies made legacy records discoverable without weakening access, ensuring granular access controls and confirming agency data would never be used to train external models
- Unstructured data made actionable: Automated ingestion and metadata extraction transformed decades of documents, emails, and attachments into a searchable context layer
- Governed, human-in-the-loop workflows: AI augments staff decisions with high transparency and source citations, while subject matter experts maintain final oversight, keeping public-service decisions accountable
The cautious approach to AI
Like many municipal governments, Northern Rockies Regional Municipality dealt with cost sensitivity, aging legacy systems that posed security concerns, and resource constraints that held back technology advancements. Robert Blain, Technological Services Manager, acknowledges these challenges: “We needed to enable collaboration with constituents and still ensure solid compliance around our digital interactions and content.”
Northern Rockies initially took a cautious approach to rolling out AI-enabled tools. Protecting confidential information and ensuring it wouldn’t be used improperly in training or prompting an AI model was paramount. In fact, Blain bluntly explains, “We issued an email to all employees that they weren’t to use AI. Period. We first needed to make sure we were in a secure environment.”
We became comfortable knowing that Box is not going to be training on our data. And Box exceeds most countries’ strict security policies.
Northern Rockies needed a solution that would address their challenges, but with a balance of control and flexibility. The agency ultimately turned to Box as the governed content layer, offering AI capability through which AI can act on authorized content. Having granular, role-based access control and comprehensive audit logging over content in Box gave Blain the confidence to responsibly deploy AI within a secure perimeter. This helped eliminate the governance risks of shadow AI and unmanaged point solutions.
He explains, “We became comfortable knowing that Box is not going to be training on our data. And Box exceeds most countries’ strict security policies.”
This gave Blain’s team the faith they needed to protect and govern the content both fed to and produced by AI. As a first use of AI, Northern Rockies tackled better ways to surface important information from massive stores of unstructured data: files, contracts, and more. Using AI, they could query data directly instead of manually sifting through the documents one by one.
For example, when a new technology arrived with complicated, unwieldy documentation, Blain was able to quickly solve a technical problem by asking Box AI to query the materials. He describes: “It came back with five documents and four solutions. Within seconds I was solving a technical problem just by using AI to query our unstructured data.”
From email archaeology to instant, governed retrieval
The confidence the team got from the first toe-dips into AI led to more and more use cases. One huge pain point for Northern Rockies (and a lot of organizations) is that sharing sensitive files and information via email is a huge vector of risk. Blain’s team partnered with San Francisco-based software company mxHERO, whose platform integrates seamlessly with Box, to automate the ingestion of agency emails and their attachments directly into Box — making it easier to find attachments, and more importantly, to apply governance rules.
This helped facilitate the handling of Freedom of Information Act (FOIA) requests. Now, when a FOIA request comes in from a constituent, managers no longer need to comb through legions of email to find information related to a specific incident. Box AI finds relevant information quickly and reliably. This automation has reduced the cost and time it takes to address FOIA requests so much that Blain expects more people will start making requests due to easier access.
Along the way, Box AI has revealed surprising insights about the kinds of questions users ask, helping Blain better focus his efforts on solving constituent problems. Additionally, automated metadata extraction across heterogeneous inputs (photos, video footage, scanned incident files. etc.) structures previously opaque data. This enables staff to pinpoint relevant records in seconds without manual sorting, maintaining full chain-of-custody and audit readiness.
Automation, but make it human-controlled
Despite the opportunities AI automation has granted Northern Rockies, Blain emphasizes that it’s still a tool to augment, not replace, human judgment. He says, “AI requires human oversight for context and the subtleties only humans understand.”
For instance, sarcasm, tone, and sincerity are difficult for AI to detect, so at Northern Rockies, humans review AI-generated content to ensure it’s appropriate and authentic.
Blain also stresses the importance of using multiple opinions and sources to validate AI results rather than relying on a single output. Ultimately, he says, “The human is in control and has the final word and ownership.” This balance preserves trust and ensures communications with constituents always feel genuine.
AI-powered HR document generation
Northern Rockies is now instituting AI transformation across its HR organization. The HR team manages numerous union agreements and employment types, from firefighters to lifeguards to public works staff, which can make HR document creation complex and time consuming.
Blain is thinking ahead of the possibility of using Box to create custom AI agents (automatic workflows based on given instructions) for various HR scenarios, such as an agent that can generate offer letters tailored to each position’s clauses and language. “Box Doc Gen could create a first draft that HR can tweak, saving hours spent pulling together documents or copying clauses incorrectly,” he describes.
Box Doc Gen creates a first draft that HR then tweaks, saving hours spent pulling together documents or copying clauses incorrectly.
This would free HR professionals to focus on strategy and employee relations rather than repetitive paperwork. Northern Rockies is working with Box Consulting to modernize such HR processes, recognizing that thousands of tasks could benefit from AI assistance.
Looking ahead, Blain sees immense value in leveraging AI to classify content and extract deeper contextual metadata across the municipality’s vast data repository. “It’s a big pool of unstructured data, and once you add structure, you really have power,” he says.
Advice for other municipalities starting their AI journey
With AI becoming mainstream and AI agents joining the workforce, smaller agencies suddenly have powerful tools to automate key workflows stemming from their unstructured data. Blain encourages other leaders to start early and small: “Make use of the Box community and GitHub. Others have gone before you, and there are examples that can walk you through the process.”
He also recommends engaging Box’s expert consultants who bring fresh perspectives and can identify key problems that insiders might miss. He encourages agency technology leaders to “do something — even if it’s small — just to get exposure.”
By starting small, building trust in secure platforms, and keeping humans firmly in control, the agency is unlocking the true potential of intelligent automation. As Blain puts it, AI is “a new superpower” that advances what an agency can do by orders of magnitude, “if it’s embraced thoughtfully.”
To learn more about Northern Rockies Regional Municipality’s use of Intelligent Content Management, watch the video.
