Thanks for subscribing to Digital Journal's What does this really mean? Follow us on LinkedIn for updates throughout the week. |
Brian Gue, manager of data science at PCL, speaks with Digital Journal CEO Chris Hogg at Edmonton Unlimited. - Photo by Jennifer Friesen, Digital Journal
On a recent trip to Edmonton, Digital Journal’s Chris Hogg attended a community coffee, hosted by Edmonton Unlimited, where he had a chance to learn more about PCL’s push to bring order to industrial project data.
PCL is working to standardize and structure the flow of information across large industrial projects. Think engineering drawings, revisions, compliance documents, and handover packages.
The goal is to standardize data early to prevent the downstream rework that slows procurement and construction.
Reading it, I couldn’t help but notice the similarity with some other conversations I’ve had recently.
For all the hype around AI, the executives and data leaders I’ve spoken with keep returning to the same issue. The underlying data has to be consistent, documented, and trustworthy.
AI ambition meets data reality
In Chris’s piece, the problem is how inconsistent data and shifting standards slow work and require people to keep cross-checking and correcting information. When project information isn’t unified, every handoff becomes manual and every tool ends up reinforcing those gaps rather than fixing them.
That same theme surfaced this week in a conversation with Geoff Nielson, SVP of global service and delivery at Info-Tech Research Group. He was direct about the stakes. Organizations that have delayed dealing with messy data and legacy systems are already behind, and catching up is getting harder. For some, it’s becoming a question of survival. (More to come from that conversation soon.)
I heard something similar in an earlier interview with Anahita Tafvizi, chief data and analytics officer at Snowflake, while reporting on why documentation debt might be holding back your AI strategy. Documentation debt builds slowly. Systems evolve. Teams change. Institutional knowledge lives in inboxes or in people’s heads.
Over time, the organization loses a shared map of its own data landscape.
Put those conversations next to PCL’s effort and the connection is straightforward. Whether it’s a construction firm managing thousands of project artifacts or an enterprise deploying generative AI, the same rule applies: garbage in, garbage out.
AI does not compensate for missing context. It scales whatever structure already exists.
Vegetables Before AI Dessert
Data quality often sounds like a technical footnote. But it shapes strategy in practice.
Early naming decisions stick. If every team labels files differently, that confusion builds.
Documentation is institutional memory. Systems only work as long as the right people are still in the room.
Governance requires ongoing attention and clear ownership.
AI shines a light on your data practices. The gaps become hard to ignore.
Across construction, enterprise IT, and AI strategy, the conversation keeps returning to the same point. The sophistication of the tool matters less than the integrity of the data it depends on.
Watercooler links
Some light reading for when your AI strategy feels confident but your data model is side-eyeing you.
A look at how new technologies are being folded into infrastructure projects, where big promises meet the slower realities of procurement, integration, and standards.
Turns out asking a chatbot “are you sure?” improves reliability about as much as asking your toddler if they definitely don’t need the bathroom before wrestling them into a snowsuit. A look at why reassurance prompts rarely fix deeper model limitations.
Microsoft researchers are experimenting with encoding data into glass designed to last thousands of years. Good to know we’ll leave behind some data for our robot overlords.
Least shocking news of the week: men building systems for men. A computing professor points to an “alpha male” culture in AI that sidelines women and reinforces existing power dynamics inside technical fields.
A reminder that rolling out new systems is the easy part. Getting humans aligned, engaged, and willing to change their behaviour is where most transformation efforts stall.
From the Digital Journal Insight Forum
The Insight Forum is Digital Journal's thought leadership platform, offering experts a dedicated space to share their perspectives with our audience across Canada, the U.S. and abroad. Members publish monthly articles showcasing industry insights and what they’re learning and seeing in their space.
A framework for assessing how executive candidates understand and manage organizational risk.
A practical overview of common cyber threats and how they typically manifest inside organizations.
An exploration of the tension leaders face between staying operationally involved and stepping back to lead strategically.
Final shots
We keep talking about AI like it’s a leap into the future.
Like many things in life, the work needed to make that leap a reality is a lot less glamorous.
Rename the files. Align the standards. Write down what the acronym actually means. Decide who owns the data. Then maybe deploy the model.
It’s hard to build a billion-dollar business on top of a folder called “Final_v3_REAL_FINAL.”
- David