Colin Stoner was hanging out with his CTO after work when the CTO made a confession: “I need you to know that I wrote zero lines of code myself today. It was all Claude.”
For Stoner, CIO at Novo Construction, this was good news: It validated months of his own dogged evangelism about AI’s potential. But it also revealed something unexpected about technology adoption: Even the most technical people need time to embrace tools that might fundamentally change their work.
There’s a tension between AI’s transformative promise and the practical human journey to adopt it. In the construction industry, where physical work meets digital transformation — and where a misread blueprint can cost millions — this is a pronounced problem.
Key takeaways:
- Novo Construction’s AI success came from tackling high-friction manual work first, using Box AI to extract and validate certificate-of-insurance data at 98% accuracy across 1,300 COIs in just 30 days
- Fast, secure AI adoption depends on a strong digital foundation, with Novo’s prior years of cloud migration and systems work making the leap possible
- Successful AI transformation requires curiosity and steady leadership, with teams using AI to expand human expertise rather than replace it
Start by changing the work people hate
When Novo Construction began exploring AI applications, Stoner took an unconventional approach. Instead of chasing flashy use cases, he walked around the office asking people a simple question: “What do you hate about your job?”
As a CIO, Stoner recalls: “It’s a fun conversation to start with someone because they’re going to look at you very skeptically.” But the strategy was deliberate. Those pain points — the tedious, repetitive tasks that drain energy and morale — are the perfect entry points for AI adoption. “Certificates of insurance were a huge one,” he remembers.
The first AI win for Novo: Extracting metadata from COIs
Every subcontractor on every project must maintain certificates of insurance (COIs) not just throughout the project but for a decade following project completion. It was a lot of compounding paperwork over time — and in the beginning, it was literal paperwork.
“We would have stacks of open envelopes with COIs sitting on the reception desk waiting to get processed,” Stoner says.
Thankfully, Novo had already made the transition to digital files in the cloud. With that in place, Novo was able to use Box AI — specifically RAG, which improves accuracy by retrieving data from external, trusted knowledge bases before generating a response. This enabled Stoner’s team to pull out the correct data from each COI, pair it with the right project code, and then “stitch it all together across 1,300 certificates of insurance in the last 30 days,” Stoner says. “It really scales well, and fast, and it’s immediate relief for some people.”
In addition to simply extracting the data, Novo is able to compare policy numbers and amounts to company standards and make sure all subcontractors have a minimum coverage level or an umbrella policy to make up the difference. Now that this automated system is up and running, Novo’s email endpoint software dumps certificate-of-insurance files into Box every hour as part of an AI workflow. Stoner describes it as “the first real AI win we had.”
Building on 20 years of digital foundation
Novo’s AI success didn’t happen overnight. The company had been building its digital infrastructure for two decades, starting with simple database extensions and evolving into Sentinel, their custom construction management platform. This patient, methodical approach to digitization created the foundation that makes today’s AI applications possible.
“Moving all our content to the cloud 13 or 14 years ago was step number one for us to have better access to information,” Stoner says.
By carefully selecting partners like Box for content management and building on that, Novo created a secure, integrated environment where AI could flourish. Stoner says, “The best you can do is lay a foundation. You have to have good groundwork, a good base to build on.”
This foundation proved critical when AI capabilities exploded onto the scene. While other construction companies scrambled to digitize paper processes and consolidate disparate systems, Novo was ready to layer intelligence on top of their existing digital infrastructure.
Imagination is a key driver in AI innovation
Perhaps the most surprising challenge in AI adoption isn’t technical; it’s imaginative.
“It takes so much imagination to be able to get the most out of these tools,” Stoner emphasizes.
This became clear when superintendents at Novo started discovering unexpected uses for AI, like using chatbots to train junior staff on reading construction schedules.
One superintendent suggested using the internal chatbot to help an assistant superintendent who was “pretty novice at scheduling.” Within two minutes of prompting, they had an interactive training tool that could explain critical path items and scheduling concepts using live project data.
The applications keep expanding. Field workers now use AI to identify wiring configurations when installing smart thermostats. Project managers use it to draft RFIs (Requests for Information) with consistent formatting and detail. Even sensitive situations benefit. One employee asked AI how to handle a subcontractor behind on payments, getting diplomatic guidance free from personal bias.
Patience as an AI leadership strategy
Leading AI transformation requires a delicate balance of change management, and that balance looks different in every company. Stoner’s approach centers on patience and demonstration. Rather than mandating AI use, he continues building tools and letting results speak for themselves.
“The sooner you learn to use it and adopt it, when those new iterations or new versions come out, you’re able to leverage whatever additional horsepower’s in there,” he advises.
This patience extends to addressing legitimate concerns. Construction work remains fundamentally physical. AI can process documents but can’t pour concrete or install drywall. Stoner frames AI as the “10 to 20% of expertise” that enhances human judgment rather than replacing it. Experienced superintendents with AI tools become “Superman on the job site,” combining decades of field experience with instant access to project knowledge.
Experienced superintendents with AI tools become “Superman on the job site,” combining decades of field experience with instant access to project knowledge.
One unexpected benefit of AI adoption has been documentation consistency. When AI writes RFIs or processes paperwork, it follows standards precisely every time.
“The level of consistency between projects, between documents that AI can produce... bless our project engineers, they’re just not going to be able to keep up,” Stoner observes.
This consistency creates a virtuous cycle. Better documentation improves AI training data, which enhances future AI performance. Each successful implementation makes the next one easier, building organizational confidence in the technology.
Beyond the three-month horizon
When asked about AI’s future in construction, Stoner laughs at the idea of making predictions. “Three to five years is eternity,” he says, noting how quickly capabilities evolve. Instead, he focuses on positioning Novo to adapt quickly to whatever comes next.
His advice for other organizations is pragmatic: build strong foundations, choose partners carefully, and don’t rush.
“You're not going to miss the boat,” he reassures, “but also be sure to be thorough and do it right.”
Most importantly, he emphasizes starting with imagination and experimentation. “The best thing you can do is just get in the weeds and play with this stuff,” he suggests, sharing how he once built a sports betting bot just to understand AI capabilities better (it went broke with fake money — twice).
As Stoner’s CTO revealed that evening over beers, transforming work with AI is a matter of giving teams the time, space, and trust to transform along with it.
Read about more construction companies using Intelligent Content Management from Box and watch the full AI-First episode below.
