Three wow moments from Summer AI Day 2026

Last month we hosted our biggest AI Day to date. More people registered than for any AI Day in nPlan's history, and despite a red weather warning turnout was extremely strong. Fortunately the offer of ice cream, a well stocked bar, and capable AC convinced plenty of folks it was worth braving the train to attend.

Now, as nPlan's Marketing Lead, my mind on the day was rather too taken up with running the event to properly absorb what was being shown on stage. It's only since watching the recording back that I've been able to appreciate just how much ground we covered, and how far our product range has come in the past 5 years.

So here, for the project professionals who were there and want a rewatch, and the ones who could not make it, are three moments from the day that noticeably changed the atmosphere in the room. I have picked the clips I think best show where project controls and delivery are heading. I would encourage you to watch the full recording to put my selections into context. Anyway, without further ado, let's dive in.

Building a $15 billion project from a few minutes of Googling

For the first time on stage, we ran a live demo of Decision Intelligence, our newest product, against a real megaproject: the $15 billion Lighthouse data centre, part of the US Stargate programme. Our Product Manager, Richard, started with nothing but publicly available documents, built a credible schedule, and then used Decision Intelligence to surface the project's biggest strategic decisions and put hard numbers against each one:

And here's nPlan CTO and Co-founder Alan Mosca summing up after Richard's demo:

This is incredible. With only publicly available information that we found through a few minutes of Googling, we were able to create an initial grounded and believable schedule, run several options through Decision Intelligence to surface a decision opportunity that could improve the cash position of the project by up to $296 million. It's a couple of hours of work.

As was hopefully clear from Richard's demo, Decision Intelligence goes well beyond schedule optimisation or crashing activities to save a few days. It works at a strategic level, weighing the real options in front of a project, the different scenarios you might pursue, and the decisions you actually need to make, then telling you what each one is worth. On Lighthouse, that meant surfacing a way to rethink the project and hit revenue milestones sooner, to the tune of a $296 million swing in cash position. The question shifts from "is my schedule efficient?" to "which future should we choose?"

The mother of agents

Next, Alan introduced Khaleesi, the new orchestration engine behind Barry, our project controls AI agent. Khaleesi can coordinate up to 200 sub-agents at once, self-directing and self-correcting towards a goal you set, without you having to spell out every step along the way. He introduced it alongside a problem every planner in the room recognised instantly: the slow slide of project controls into what he called "data janitorship":

For years, project controls has felt a bit like an administrative trap. It's less about leadership and more about data janitorship… I think that's a massive waste of human potential. Barry and its new Khaleesi engine can change all that, moving beyond basic automation to deliver a true digital workforce.

Barry, live in the room

We closed out the keynote with a demo of something genuinely impressive which is in the early stages of development: Barry Live. Think Barry sitting in on a meeting - not to take notes, but to do the work. The team briefed it out loud, and Barry analysed a Decision Intelligence scenario, named the biggest risks, built a leadership-ready deck, and spun up an interactive economics dashboard, adapting on the fly when the team changed direction mid-conversation.

Good, Damian. Ready when you are. I can see that you're trying to lock in the [data centre] project.

Alan was careful to frame this one honestly: it's a research preview, with a few bloopers left in and some interactions still to iron out. But even in preview, it shows the shift we've been working towards. Until now, getting an answer meant going to a tool with your question. Barry Live flips that around. The platform comes to you, in the room, while the conversation is still live, bringing the collective intelligence of every project we've ever forecast on with it, plus the tooling to analyse, model and act on it there and then.

Watch the whole thing

Those are my three, but there was plenty more on the day. If you were there, relive it. If you missed it, here's your chance. The full recording is available to watch now.

Watch the recording →

AI Day returns in 2027. Watch this space.