AI massing agents generate feasible building masses inside a buildable envelope and handle the technical decisions that follow. In Snaptrude, two agents do this work in the Feasibility phase: Explore Massing generates feasible masses inside the buildable envelope and surfaces GFA, floor count, and efficiency metrics; Place Cores positions elevator, stair, and service shaft cores inside an approved massing while validating lift capacity and egress compliance.

Used together, they change what massing exploration looks like in practice. Instead of committing to one or two schemes because testing more is too slow, you can iterate on feasible options before anything is locked in.

Why does massing feel like a commitment you're making too early?

Every architect knows the feeling. You've done the site analysis, you have the brief, and now it's time to start massing. You sketch one scheme, maybe two. You pick the stronger one and move forward. Three weeks later, you're deep into schematic design on a massing you chose in the first two days, and you're no longer sure it's the right one.

This is not a failure of process. It's a consequence of how long manual massing exploration takes.

To properly compare massing configurations, you need to model each one, check it against the buildable envelope, get a rough sense of floor plate efficiency, work out where the core goes, and assess how each scheme affects the floors above. That's a day of work per scheme, conservatively. On a real project timeline, with a client presentation on Friday, you test two and pick the better one.

The problem is that massing is where the most consequential design decisions happen. Floor plate shape determines everything: natural light penetration, structural efficiency, construction cost per square metre, the experience of every room on every floor. A massing decision made with limited exploration tends to calcify. Teams build on top of it, and by the time someone questions whether it was the right call, the cost of changing it is too high.

Here's what we noticed: the architects who produce the best buildings are not always the most talented at individual design moves. They're the ones who test the most options before committing. More exploration, better outcomes. The constraint has always been time.

AI massing exploration in architecture addresses that constraint directly.

Key Takeaway: Massing is where the most consequential decisions happen, but it's also where architects have the least time to explore. That tension is what these agents are built to resolve.

What do the Explore Massing and Place Cores agents actually do?

Explore Massing

The Explore Massing agent generates feasible building masses inside the buildable envelope produced by the Buildable Envelope agent (which builds the maximum buildable 3D envelope from setback, coverage, and height rules). It surfaces three metrics for each mass: GFA, floor count, and efficiency. These metrics give you a basis for comparing options on the numbers that matter, before you invest time in developing any of them.

The agent supports approve-or-reprompt iteration. If a mass isn't right, reprompt the agent for a different option. Once you find a mass you're confident in, approve it, and the model is committed to the canvas.

This is a different rhythm from manual massing. You're not running the agent once, taking what you get, and moving on. You're using it as an iteration loop, refining toward the right mass with each prompt, with metrics visible alongside each option to ground the comparison.

To run Explore Massing, you need an imported site and an approved buildable envelope. Both come from upstream agents (Analyze Site and Buildable Envelope), and both must be approved before Explore Massing can run.

Place Cores

Once you've approved a massing, Place Cores positions elevator, stair, and service shaft cores within it. The agent uses the massing envelope and floor count as inputs, and validates the placement against two criteria: lift capacity and egress compliance.

What this means in practice: the agent doesn't just put a core in a default position. It checks whether the placement satisfies vertical transport demand for the building height, and whether emergency egress travel distances meet code requirements. The validation is the value here. You're not just getting a core placement; you're getting one that's been checked against the technical constraints that matter most.

Place Cores requires an existing massing on the canvas. It runs after Explore Massing and feeds into the program-packing stage that follows.

Key Takeaway: Explore Massing generates feasible masses inside the buildable envelope and surfaces GFA, floor count, and efficiency metrics for each. Place Cores positions cores inside a chosen massing and validates lift capacity and egress compliance.

How do you use these agents together in a real workflow?

The sequence matters, and there's a strict dependency chain you need to respect.

Start by importing your site. Both agents require an imported site via the Load Site flow. Without it, neither agent will run.

Run Analyze Site. The agent reads zoning rules, setbacks, FAR, and height limits, and produces a reviewable summary. Approve the output before continuing.

Run Buildable Envelope. This agent builds the maximum buildable 3D envelope from the setback, coverage, and height rules surfaced by Analyze Site. The envelope it produces is what Explore Massing works inside.

Run Explore Massing within that envelope. Review the masses it generates. Check the GFA, floor count, and efficiency metrics it surfaces. If the first option isn't right, reprompt. Iterate until you find a mass that fits the brief and the site. Then approve it.

Run Place Cores on the approved massing. The agent positions elevator, stair, and service shaft cores, and validates the placement against lift capacity and egress compliance. Adjust if the validation flags something specific to your project.

Move into programming and packing. Generate Program (which has no prerequisites) creates departments and area allocations. Stack Program distributes them across the storeys. Pack Program then runs adjacency analysis and packs the program into each floor of your approved massing.

The whole sequence, from imported site to packed program, can run in a single session. The agents are accessed from the sparkle agent launcher in the bottom command bar, which opens the Canvas Agent Panel on the left side of the canvas.

From here, the workflow connects to Snaptrude's programming agents, which take your massing forward into department-level and room-level planning.

Key Takeaway: Massing agents sit inside a strict dependency chain. Site import, Analyze Site, Buildable Envelope, Explore Massing, Place Cores. Respecting the order is what lets each agent do its job.

What are the limits of AI massing exploration?

Three things worth knowing.

Explore Massing generates feasible masses, not architecturally resolved ones. The masses it produces fit inside your buildable envelope and surface useful metrics. They don't account for view corridors, neighbourhood character, structural intent, or the particular logic of your client's brief. Those are design judgments you bring. The agent narrows the field; it doesn't make the call.

Place Cores validates lift capacity and egress compliance, but doesn't replace a fire engineering review. The validation gives you confidence in the core placement at the feasibility stage. A formal life safety review still belongs to your engineers later in the process. Treat the agent's validation as a strong starting point, not a finished compliance document.

The dependency chain is real. You can't shortcut Explore Massing without an approved buildable envelope. You can't run Place Cores without an existing massing. If you're trying to use these agents on an existing project that wasn't set up through the upstream agents, you'll need to bring those inputs into the project first. The chain exists because each agent depends on the output of the one before it.

These are real constraints, not caveats. Working within them, Explore Massing and Place Cores handle the geometric and technical heavy lifting of early massing so you can focus on the design decisions those geometries enable.

Key Takeaway: These agents operate inside a defined chain and produce feasibility-stage outputs, not final design or compliance answers. Used in sequence, they handle the technical work that early massing depends on.

More options, better buildings

The case for massing exploration is simple: the more configurations you test before committing, the more likely you are to find the right one. Most architects already know this. The constraint has been that testing takes time, and time runs out before the options are exhausted.

Explore Massing and Place Cores are built around that constraint. They don't make the design decision. They make sure you've seen enough options that your decision is well-founded, and they handle the technical validation that the decision depends on.

If you're mid-project and wondering whether your current massing is the right one, it's worth running the agents on the same brief and seeing what comes back. Sometimes it confirms what you already have. Sometimes it surfaces something you didn't consider.

Either way, you'll know you made the call with more information than most architects get.

Try Explore Massing on your next project in Snaptrude or read how these agents connect to the space planning stage when massing is resolved. For the full picture of how site, massing, and program connect in one workflow, see the 2.5-hour RFP workflow.

We built these agents because the best massing decisions we saw in early projects came from teams who had the time to explore properly. We wanted every team to have that time, regardless of deadline pressure. Whether it works that way in practice is something you'll find out quickly.