Introduction
In the rapidly evolving world of Architecture, Engineering, and Construction (AEC), leveraging cutting-edge technologies to streamline processes has become essential. Generative layout automation is one such innovation that is redefining site masterplanning. By integrating Forma and Speckle, architects and planners can harness the power of artificial intelligence and data-driven decision-making to enhance the planning phase and foster creativity.
The pressure on design teams has intensified considerably over the past decade. Clients expect more scheme options in shorter timeframes, local authorities demand comprehensive environmental assessments from the outset, and sustainability targets grow more exacting with each revision of planning policy. Against this backdrop, the traditional workflow — a lead architect sketching massing options, circulating PDFs, collecting comments by email, and manually updating a single Revit model — is simply no longer fit for purpose on anything above a small residential project.
Generative layout automation addresses this gap directly. Rather than replacing the designer's judgement, it amplifies their capacity to explore, evaluate, and refine — turning what was once a linear, laborious process into a parallel, evidence-based one.
What is Generative Layout Automation?
Generative layout automation refers to the process of using algorithms to generate design alternatives swiftly. This technology enables designers to set constraints and goals, and the software generates numerous iterations that meet these criteria. This is particularly useful in site masterplanning, where variables such as zoning laws, environmental conditions, and client requirements must be optimally balanced.
At its core, generative design treats the design brief as a set of mathematical objectives. Floor area ratio (FAR) targets, daylight angle requirements, setback distances from boundaries, fire engine turning circles, noise contours from nearby roads — each of these can be encoded as a constraint or an optimisation target. The algorithm then searches the solution space, producing dozens or hundreds of viable massing configurations that satisfy every hard constraint whilst pushing towards the soft objectives as far as possible.
The distinction between hard and soft constraints is worth dwelling on. Hard constraints are non-negotiable: a building cannot legally exceed a given height in a conservation area, or must provide a minimum quantum of affordable housing. Soft constraints are preferences that the design team ranks and weights — maximise open space, minimise overshadowing of neighbouring gardens, reduce embodied carbon in the structural frame. Generative tools allow these weightings to be adjusted interactively, so a planning consultant and a sustainability engineer can each see the scheme that most favours their respective priorities before a single compromise is agreed.
The approach draws on techniques from parametric modelling, evolutionary computation, and — more recently — machine learning. Some platforms use genetic algorithms that mimic natural selection, mutating and recombining candidate layouts over successive generations. Others employ gradient-based methods that iteratively nudge geometry in the direction that improves a given metric. The practical result for the design team is the same: a rich set of explored options, each accompanied by quantitative performance data, delivered in a fraction of the time manual iteration would require.
The Role of Forma and Speckle in Site Masterplans
Forma's Contribution
Forma, developed by Autodesk, is a cloud-native early-stage design platform that embeds environmental analysis directly into the massing and layout workflow. Where traditional practice required a separate specialist engagement — a daylight consultant running IES VE, a microclimate engineer running CFD in ANSYS Fluent — Forma surfaces equivalent intelligence inside the design environment itself, updated continuously as geometry changes.
In practical terms, Forma ingests site context from GIS sources (OpenStreetMap, national cadastral datasets, local authority 3D city models) and uses this data to run real-time solar radiation analysis, wind speed simulations based on Lawson criteria, noise propagation modelling, and pedestrian comfort assessments. The results are not static reports but live visual overlays: a heatmap of annual sunlight hours on every facade surface, animated wind vectors through the public realm, contours of noise exposure across the ground plane.
For a masterplan project, Forma's generative massing tools allow the team to define a site boundary, input planning constraints such as height limits and required street setbacks, and then use the platform's AI-assisted layout engine to propose configurations that optimise for a chosen mix of objectives. A firm working on a mixed-use regeneration scheme in Birmingham, for example, used Forma to evaluate 47 distinct block configurations in a single afternoon — a task that would previously have occupied a junior architect for two weeks of manual modelling. The shortlisted schemes were then presented to the local planning authority with full environmental data attached, significantly accelerating the pre-application process.
Forma also integrates directly with Revit through the Forma Add-in, allowing teams to promote a preferred early-stage massing into a detailed BIM environment without re-modelling from scratch. Geometry, site context, and analysis results travel with the model, maintaining continuity from concept through technical design.
Speckle's Advantage
Speckle offers an open-source platform that enhances real-time collaboration among architects, engineers, and stakeholders. By facilitating seamless data exchange across different software packages and between distributed team members, Speckle ensures that everyone works from the same current model throughout the design process.
Speckle's architecture is built around the concept of streams, branches, and commits — a version-control paradigm familiar to software developers but relatively new to AEC. Each discipline contributes to its own branch: the structural engineer pushes updated column grids and slab outlines; the MEP consultant pushes revised plant room allocations; the landscape architect pushes updated ground level topography. The Speckle server merges these contributions and tracks every change, making it straightforward to compare the current state with any previous commit or to roll back a single discipline's work without disturbing the others.
Connectors are available for Revit, Rhino, Grasshopper, AutoCAD, Civil 3D, Blender, Unity, and several other applications, which means data rarely needs to be re-exported or re-imported manually. A parametric script running in Grasshopper on a landscape architect's machine in Edinburgh can push updated ground modelling directly to a Speckle stream; a Revit user in Nairobi sees the updated terrain in their model within seconds, without any file transfer or manual coordination.
Consider a live infrastructure project where a highways consultant needs to update road alignment geometry in response to a revised junction design. In a traditional workflow, this involves exporting a DWG from Civil 3D, emailing it to the architect, waiting for them to import and check it, then manually adjusting the masterplan layout. With Speckle, the updated alignment pushes to a shared stream automatically on each Civil 3D save, and the downstream Revit model references the stream natively — the architect simply refreshes the Speckle receiver and sees the current geometry without any file management overhead.
A Step-by-Step Implementation Workflow
Step 1: Define Goals and Constraints
Begin by translating the project brief and planning policy context into a structured parameter set. At minimum this should include: site boundary geometry (as a georeferenced polygon), height limit envelopes by zone, minimum daylight standards (typically BRE 209 targets for residential), floor area requirements by use class, and any protected views or heritage constraints. Forma's constraint input panels accept these directly; more complex regulatory envelopes can be imported as 3D solids from CAD sources.
It is equally important at this stage to define the optimisation hierarchy — which objectives are hard constraints and which are soft targets to be weighted. Involve the planning consultant and sustainability lead in this conversation. Their input directly shapes which schemes the algorithm surfaces and can prevent wasted iterations on options that are technically viable but commercially or politically undeliverable.
Step 2: Generate Initial Layout Ideas
Activate Forma's generative massing tools with the constraints defined above. Depending on the site's complexity, the platform will typically return between 20 and 60 candidate configurations within a few minutes. Each configuration is immediately analysed for solar access, wind comfort, and gross internal area — the three metrics most consistently demanded at pre-application stage.
Review the initial results critically. The algorithm is not aware of every nuance the design team holds — it does not know that a particular corner of the site has been identified as a future tram stop, or that the client has a strong preference for a central public courtyard. Use this first pass to calibrate the constraint set: tighten parameters where the algorithm is finding solutions the team would immediately reject, and loosen parameters where the solution space appears artificially constrained.
Step 3: Collaborate and Adjust
Export the shortlisted massing options from Forma and push them to a Speckle stream using the Forma-to-Speckle connector or via Rhino as an intermediate. Organise the stream with a branch per scheme option, so the multidisciplinary team can navigate between alternatives without confusion.
Share the stream with all consultants and with the client. Encourage asynchronous review: Speckle's web viewer allows anyone with a browser to examine the 3D models, inspect element properties, and leave comments anchored to specific geometry. A structural engineer can flag a scheme where the column grid implied by the proposed facade module is incompatible with an efficient flat-plate structure; the planning consultant can annotate a massing that encroaches on a protected view corridor the algorithm did not model.
Consolidate feedback and update the Forma model accordingly. Where feedback reveals a gap in the original constraint set, add the missing parameter and regenerate — the turnaround is measured in minutes rather than days.
Step 4: Evaluate and Optimise
Return to Forma for a detailed performance pass on the revised preferred option or the two to three schemes that remain in contention. At this stage, move beyond basic solar access and look at more granular metrics: facade solar gain by orientation (relevant to HVAC sizing), shadow impact on neighbouring residential windows using BRE 209 vertical sky component calculations, and pedestrian wind comfort at key entrances and in the public realm using Lawson criteria classification.
If the project has a net zero carbon commitment, use Forma's embodied carbon estimator to compare structural material quantities across the competing schemes. Even at massing stage, the choice between a compact courtyard block and a looser pavilion arrangement can produce a 10–15% difference in structural material volume, a meaningful early indicator of embodied carbon performance.
Step 5: Finalise the Masterplan
After thorough analysis and collaborative revision, arrive at a preferred scheme that satisfies all hard constraints, performs well against the weighted soft targets, and has been validated by the full consultant team. At this point, promote the Forma model to Revit using the Autodesk Forma Add-in. The promoted model arrives as a georeferenced massing with site context attached — a clean starting point for the detailed BIM model that will carry the project through planning and technical design.
Archive the full generative study — all candidate schemes, their performance data, and the evolution of the constraint set — in Speckle. This record has practical value at planning stage: being able to demonstrate to an inspector that dozens of massing options were evaluated against specific environmental criteria, and that the preferred scheme demonstrably outperforms the alternatives, is a compelling argument in support of a planning application.
Case Studies
Urban Quarter Regeneration, West Midlands
A regional developer engaged an AEC consultancy to masterplan a 4.2-hectare brownfield site on the edge of Birmingham city centre for a mixed-use scheme comprising 650 residential units, 3,500 sq m of ground-floor retail, and a new public square. The planning context was complex: a grade II listed textile mill sat immediately to the north, a busy arterial road ran along the southern boundary, and the local plan required a minimum of 30% affordable housing with no cross-subsidy reduction.
Using Forma, the team ran 53 massing configurations in the first two working days. The platform's solar analysis immediately eliminated 18 configurations that would have cast the public square into significant shadow during the winter months — a requirement the client had stipulated as non-negotiable. The remaining 35 were pushed to a Speckle stream and reviewed by the heritage consultant, the acoustic engineer (assessing the road noise impact on residential facades), and the structural engineer in a single two-hour online session.
Within one week the team had a shortlist of three preferred options, each with full environmental data and a structural feasibility note. The pre-application meeting with the planning authority took place in week four of the project — a programme that the practice director estimated would have taken three to four months under the firm's previous workflow. The authority responded positively to the evidence-based presentation of the design evolution, and a planning application based on the selected scheme was submitted without a single request for additional environmental studies.
University Campus Extension, Scotland
A Scottish university commissioned a masterplan for a new academic campus extension on a steeply sloping site adjacent to an existing listed building complex. The project required careful management of view corridors from the listed buildings and compliance with the university's net zero 2035 commitment.
The design team used Forma to model the site topography from LiDAR survey data and then ran generative layout options constrained by the protected view cones. Speckle connected the landscape architect's Civil 3D terrain model, the architect's Revit massing studies, and the structural engineer's early foundation assessments in a single coordinated stream. The heritage consultant reviewed scheme options directly in the Speckle web viewer from their office in Edinburgh without ever needing to install specialist software.
The embodied carbon analysis in Forma guided the team towards a scheme with a compact footprint and shared structural cores between adjacent buildings, reducing the estimated structural steel tonnage by 22% compared with the initial loose-pavilion arrangement preferred on purely aesthetic grounds.
Measuring the Business Impact
The efficiency gains from this integrated workflow are quantifiable, and it is worth understanding what the numbers actually mean for practice economics and project delivery.
Iteration speed. A conventional masterplan study that generates four to six massing options over four weeks of senior architect time can be replicated — with ten times the number of options and full environmental data — in three to five working days using Forma. This does not mean that the senior architect's time is displaced; it means their time is redirected from repetitive modelling to evaluative judgement, where their expertise adds the greatest value.
Coordination overhead. Speckle eliminates a substantial proportion of the inter-disciplinary coordination overhead that typically consumes 15–20% of a project's design fee. Internal studies by practices adopting the platform report a 30–40% reduction in RFI (request for information) traffic during the concept and schematic design phases, with a corresponding reduction in the rework triggered by late detection of coordination clashes.
Planning success rates. Practices that present generative design studies at pre-application stage report noticeably higher rates of receiving constructive planning feedback and lower rates of refusal on masterplan applications. The ability to demonstrate that the preferred scheme has been tested against a wide range of alternatives — and that it demonstrably outperforms them on the authority's stated criteria — shifts the nature of the dialogue from adversarial to collaborative.
Client confidence. For developer clients, the ability to see quantitative performance data alongside scheme options accelerates internal approval processes. Rather than debating which massing option "feels right", decision-makers can compare schemes against commercial metrics (GIA, unit mix, car parking quantum) and sustainability metrics (solar access, embodied carbon) simultaneously, making investment decisions on the basis of evidence rather than instinct.
Best Practices and Common Pitfalls
The workflow described above delivers its full benefits only when certain disciplines are observed. Teams adopting generative layout automation for the first time frequently encounter the same set of challenges.
Garbage in, garbage out. The quality of Forma's output depends entirely on the accuracy of the constraint set. If the site boundary polygon is incorrect, if height limit envelopes are over-simplified, or if the protected view cones are not modelled accurately, the algorithm will optimise towards the wrong target. Invest time upfront in validating inputs against the planning policy documents and the project brief before running any generative studies.
Avoid optimising too early. The generative phase should be exploratory, not immediately convergent. Resist the temptation to narrow to a single preferred option after the first round of results. The value of the tool lies in the breadth of the solution space it can explore; compressing that space prematurely sacrifices the insight that broader exploration would have delivered.
Version discipline in Speckle. The branch-and-commit model is powerful, but only if the team maintains consistent naming conventions and commit messages. Establish a simple naming protocol at project kick-off — branch names should identify the discipline and the design stage, commit messages should describe the specific change — and enforce it throughout. Without this discipline, the stream becomes difficult to navigate and the audit trail loses its value.
Combine automated analysis with professional judgement. Forma's environmental simulations are powerful early-stage tools, not substitutes for detailed specialist reports. The solar analysis results are approximate; they do not replace a BRE 209 daylight and sunlight assessment for planning. Use the automated outputs to guide design decisions and eliminate poor-performing options, but commission specialist assessments for the preferred scheme before submitting a planning application.
Engage stakeholders early with the Speckle viewer. Community consultation and client review are consistently more productive when participants can interact with 3D models in a browser rather than viewing 2D plans on a screen. Introducing the Speckle viewer to non-technical stakeholders early in the process, and taking the time to explain what they are looking at, pays dividends throughout the project.
Conclusion
Generative layout automation with Forma and Speckle represents a fundamental shift in how site masterplanning is practised, not merely an incremental improvement to existing workflows. The combination of AI-assisted environmental analysis, rapid iterative massing, and open data collaboration removes the bottlenecks that have historically made masterplanning slow, expensive, and resistant to the kind of thorough option-testing that produces truly excellent places.
The practices and developers that are adopting this workflow today are gaining a structural advantage: they are delivering more rigorous design studies in shorter timeframes, building stronger cases for planning approval, and making better-informed decisions at the stage when those decisions cost the least to change. As the tools mature and the data infrastructure around them improves — richer GIS datasets, more accurate environmental simulation, tighter integration with cost and carbon modelling platforms — the competitive gap between practices that embrace generative methods and those that do not will only widen.
At Adyantrix, we work directly with architectural practices, developers, and public sector clients to implement and optimise generative design workflows tailored to their project types and organisational structures. Whether the goal is to accelerate a complex mixed-use masterplan, build an internal capability for environmental analysis-led design, or integrate Speckle-based coordination into an existing Revit practice, our BIM automation and architectural BIM services teams bring the technical depth and project experience to make the transition efficient and durable. If you are ready to move beyond the constraints of conventional masterplanning workflows, we should talk.
Speak with our BIM Automation team at Adyantrix to find out how we can support your next project.
