Project Data Management is Messy
If you’re a project manager, you’d know what I’m talking about.
Your team spends hours each week wrestling with spreadsheets. Different teams keep their own database or workbook. You end up chasing down versions, cleaning up duplicates, and trying to piece together a coherent picture.
By the time you finally reconcile all that data, you’re on the clock again to compile KPIs and projections for your monthly executive report.
According to an interview with MACE in the UK, contractors devote up to 50% of their resources to litigation and administration—much of it capturing, organizing, and verifying data throughout a project’s lifecycle. In my own experience, our engineering team spends up to 15% of their time on data entry and report writing that, let’s face it, few ever read.
All this manual drudgery not only devours the bulk of your team’s time and budget, it also undermines data quality. Gartner estimates that poor data alone costs industries over $14 million per year on average.
Here comes LLM
We believe that, a couple of years from now, AI and LLM will eliminate manual data entry, data query and data administration for most of us.
This is how it’s going to work.
Say you are a general contractor managing a construction project. At the end of each day, your site team will issue a report, in natural language, to the AI agent. The AI then logs, categorises and summarises the reports. Everything is neatly organised, tracked and timestamped.
Together with the AI, the project manager and the project team will review this to assess progress, issues and risks. You check what’s accurate, because “done” on site isn’t always done on paper. Once everyone is happy, the AI will update the records. It applies the approved changes to the project plan, tracker, issue log—whatever’s needed—and logs everything for full auditability.
Issues and risks are captured and documented. Supporting files, such as photos, scanned documents, email threads, are stored in a shared repository and linked to the database tables. No more vital information lost on someone’s laptop or buried in an inbox.
Querying that data is just as simple. Want to know which tasks are behind schedule? The AI will tell you—and show you. Need a list of all outstanding issues? Just ask. It’s an embarrassingly trivial task for the AI to query a database, and it can generate query expressions with ease and precision. It’s a lot less trivial for a project manager to trawl through dozens of spreadsheets and project plans.
It doesn’t stop at structured data. The AI can search through reports, letters, and notices to make sure nothing gets missed. Need a window analysis? Ask for the report. No more trawling through email chains to figure out who said what—or when. The AI handles it almost instantly, with minimal manual intervention.
Single Source of Truth Done Right
I’ll be honest, I really don’t like the phrase Single Source of Truth. Not because it’s a bad idea, but because it’s been thrown around for decades and rarely delivered. The reason we haven’t achieved it isn’t technical, it’s human. We’re just not great at managing data.
Excel is an amazing tool and I love it! But that’s a double edge sword. Most project teams have spreadsheets coming out of their ears. Before you know, you have ten versions of the same information floating around, riddled with duplication, inconsistencies, and errors. Teams re-enter the same data again and again. And the project manager? Stuck stitching it all together, just to get a halfway accurate picture of what’s actually happening.
But when AI handles the data entry and admin, consistently, automatically, and without bias, that elusive single source of truth stops being a slogan but a reality.
Project Green Lantern
This is why we are building Project Green Lantern, an end-to-end AI platform to redefine project data management. We have just completed our AI agent v.0.2.
If you would like to know more about what we are doing, please reach out to us on [email protected].
