nPlan AI Day Winter 2026: The Project Controls Loop Just Got a Whole Lot Faster

We attended nPlan's AI Day in London. Here's what happened — and why the project controls profession should be paying very close attention.

A packed room in London. Standing room only. People turned away at the door because they'd hit the fire limit. That's not what you expect at a project controls tech event on a grey February day but that's exactly what happened at nPlan AI Day Winter 2026.

We were there. And honestly? It was one of the most impressive live product demonstrations we've seen in the AI-for-projects space.

If you missed it, the full recording is available to watch here. We'd strongly recommend carving out the time.

What is nPlan AI Day?

For those unfamiliar, nPlan has been running these AI Day events twice a year, a chance for their team to showcase what they've been building, announce product updates, and give the industry a look at where AI in project controls is genuinely heading. Not slide-deck promises. Real, working products. As Alan Mosca, nPlan's CTO, was keen to point out: "What you saw today was zero mock-ups. Everything was real."

This edition was themed around The Project Controls Loop, the idea that scheduling, risk analysis, cost management, reporting, and corrective action shouldn't be disconnected steps done by different teams in different tools over weeks. They should be a continuous, interconnected cycle. And AI can be the thing that finally makes that possible.

Project Super Intelligence — Not the Sci-Fi Kind

Alan Mosca kicked things off by introducing what nPlan has been calling Project Super Intelligence. Before anyone panics, this isn't the Terminator. It's a framework built around four principles:

• Hyperconnected — it can talk to everything. Your schedule, your cost data, your documents, your RFIs in Procore, your SharePoint files. All of it, in one place.

• Proactive — it knows what it should do before you tell it. Think of it as moving from reactive reporting to anticipatory insight.

• Preemptive — it doesn't just know what to do, it knows what might happen and gets in front of it. Risk management that actually manages risk, rather than just documenting it after the fact.

• Scaled — across speed, volume, and the fractures of complex project data. Not one project at a time. Portfolios. Programmes. Simultaneously.

• These aren't just nice-sounding principles. Everything nPlan demonstrated on stage was built to deliver on them.

Schedule Studio: From Blank Page to Execution Schedule in Minutes

The headline demo was extraordinary. Leonie (nPlan's Head of Product) walked the audience through generating a full schedule for the Chicago Red Line Extension, a real, $5 billion+ transit mega-project currently in pre-construction.

Starting with nothing but a scoping report and an alignment map, Schedule Studio:

• Extracted project-specific context (federal and state permitting requirements, geographic locations, funding sources)

• Generated a detailed WBS that reflected the actual complexity of the project

• Produced a logically-linked, duration-estimated schedule — with explainable reasoning for every duration

• Expanded high-level tasks into Level 2 and Level 3 detail at the click of a button

The whole thing took about 20 minutes. For context, this kind of work traditionally takes weeks of a planning team locked in a room.

And the new pricing? Starting at $250 per user per month. That's a statement of intent about accessibility.

Insights and Risk: The Probabilistic Critical Path

Once you've got a schedule, nPlan's Insights product forecasts how risk flows through it — powered by data from over 750,000 real project schedules. The driving paths visualisation shows the probabilistic critical path, highlighting which activities are genuinely driving delay and cost exposure.

For the Chicago Red Line demo, it flagged things like unexpected weather impacts on earthworks (Chicago isn't called the Windy City for nothing) and safety certification bottlenecks near project completion — exactly the kind of late-stage risk that catches owners off guard.

But here's where it gets really interesting. They've added a "Fix My Schedule" button that uses Schedule Studio's AI to automatically resolve quality issues expanding over-summarised activities, fixing logic gaps — without leaving the analytics environment. No more bouncing between scheduling software and risk tools.

Barry: The Agent That's Learning to Think in Workflows

nPlan's agentic platform, Barry, has been around since June 2023. But the updates announced at AI Day represent a genuine step-change.

The big revelation: nPlan has been studying how users actually interact with Barry, and they've discovered that people increasingly try to accomplish entire workflows in a single prompt. Things like: "Read the contract, check the delay forecast against it, identify mitigation strategies, evaluate them, and create a board presentation."

When they first measured Barry's ability to handle these long-chain workflows in one shot, the success rate was 4%. Today, it's 43%. That's a ten-fold improvement, and it's climbing.

They've packaged the most common successful workflows into Skills — templated, repeatable prompts that users can trigger with a click. Think of them as macros for project intelligence.

Other Barry upgrades include long-term memory (it remembers context across conversations), sub-agents for code writing and data extraction, deeper integration between Insights and Portfolio, and — perhaps most significantly — integrations with Procore, SharePoint, Power BI, and email. Barry can now pull an RFI from Procore, analyse it against your forecast, generate a response, post it back to Procore, and create a presentation — all from a single prompt.

Sizewell C: Real-World Validation

The event also featured a fireside chat with Tommy Clarke (Head of Programme Controls) and Carolyn Mayoh (PMO Director) from Sizewell C, one of the largest and most complex infrastructure projects in Europe. Hearing practitioners from a nuclear mega-project discuss how they're using AI-driven project controls wasn't just interesting, it was validating. This isn't theoretical anymore. It's operational on some of the most scrutinised projects in the world.

Our Conversation with Alan Mosca

We also sat down with Alan Mosca for a deeper conversation on the Project Flux podcast in a collaboration with Gen AI podcast. If the AI Day keynote gave you the "what," our podcast episode gives you the "why." Alan's perspective is refreshingly honest — nPlan was born out of AI research in 2017, nearly nine years of continuous R&D, and he's clear-eyed about both the potential and the limitations of AI in project controls. It's a conversation every project professional should hear.

Why This Matters

Here's the thing that stuck with us walking out of that room: this wasn't a vision deck. Every single thing demonstrated on stage was a working product. The schedule generation, the risk analysis, the agent workflows, the integrations — all real.

The project controls loop that Alan described — plan, forecast, analyse, act, re-plan — has always been the aspiration. But the tools, the time, and the data fragmentation have made it practically impossible to execute at speed. What nPlan showed is that the technology to close that loop now exists.

nPlan's next AI Day is on 25th June 2026. Based on what we saw in February, we'd recommend getting that in the diary now.

Watch the full nPlan AI Day Winter 2026 recording: nplan.io/events/nplan-ai-day-winter-2026-the-project-controls-loop

Listen to our conversation with Alan Mosca on the Project Flux podcast: Available now on Spotify, Apple Podcasts, and all major platforms.

Subscribe to the Project Flux newsletter for weekly AI and project delivery insights: projectflux.beehiiv.com