Skip to content
GULSHAD ANSARIBIM STUDIO

Coming soon

We're working on this.

This part of the studio is still being built. We're putting the finishing touches on it and it'll be here soon.

Back to home

AI × BIM × AUTOMATION

Solving Real Construction Problems with AI, BIM & Automation.

Construction workflows are changing. AI is moving beyond chat — into the tools, models and processes that build the physical world.

From Revit and Dynamo to AI agents and MCP, I build practical systems that connect intelligence with real construction workflows.

Gulshad Ansari · BIM Automation Specialist · Based in Dubai, UAE

Explore the BIM ToolkitHow I Build

How the work flows

Someone states what the project actually needs, in their own words.

Construction is still too manual.

A modern project produces an enormous amount of structured information. Most of the workflow around it still depends on manual interpretation and repetitive execution.

  • Drawings

    Redrawn, re-issued and reinterpreted across every revision.

  • Models

    Rebuilt by hand from information that already exists elsewhere.

  • Specifications

    Written once, then reconciled against the model repeatedly.

  • Schedules

    Extracted, reformatted and checked manually.

  • Coordination

    Clashes and gaps found late, by eye, across disciplines.

  • Revisions

    Every change rippling outward through documents by hand.

  • Repetitive tasks

    The same clicks repeated across hundreds of elements.

  • Project information

    Rich, structured, and largely underused.

The data already exists. The bottleneck is everything a person still has to do by hand to move it from one place to the next.

From Software Tools to Intelligent Workflows

The tools are not the problem. The problem is how much of the thinking and the execution still has to pass through a person one click at a time.

How it works today

  1. Human
  2. Software
  3. Manual decisions
  4. Manual execution
  5. Result

Where it is going

  1. Human intent
  2. AI
  3. Understand
  4. Reason
  5. BIM / CAD
  6. Execute
  7. Verify
  8. Result

Real Problems. Working Solutions.

Each of these started as something slow and repetitive on a live project, not as an idea looking for a use.

CAD → Revit

In the Toolkit

Rebuilding a model that already exists on paper

ProblemLegacy CAD drawings carry the geometry a project needs, but that geometry gets redrawn by hand as native Revit elements.

  1. BeforeSomeone opens the CAD drawing next to an empty Revit view and starts redrawing it.
  2. AI + BIMLayers, linework and block structure are read as geometry with meaning attached, not as anonymous lines.
  3. AutomationNative Revit elements are generated from that geometry, and drawing data is written onto their parameters.
  4. VerifyThe modeller checks the generated elements against the drawing before the model is relied on.
See it in the Toolkit

HVAC

In the Toolkit

Ductwork and terminals from the source drawing

ProblemDuct runs, fittings and air terminals get placed one element at a time, tracing over a CAD background.

  1. BeforeDuct runs, fittings and terminals are placed one at a time over a CAD background.
  2. AI + BIMCentrelines and terminal positions are read from the drawing and matched to the spaces they serve.
  3. AutomationConnected ductwork and terminals are generated, and airflow values are assigned from the same source data.
  4. VerifyThe engineer reviews routing and airflow before it goes into coordination.
See it in the Toolkit

AI × Revit

Exploration

Describing the outcome instead of the steps

ProblemAutomation only helps when you already know which tool to run, on which elements, in which order.

  1. BeforeAutomation only helps if you already know which tool to run, on what, in which order.
  2. AI + BIMA stated intent is interpreted against the model, so the model itself is the context.
  3. AutomationThat intent is translated into the specific model operations which would satisfy it.
  4. VerifyNothing is applied blindly — the proposed operations stay a proposal until accepted.
Discuss this work

MCP

Exploration

Giving an agent real access to the authoring tool

ProblemAn AI that cannot read or change the model can only ever advise from the outside.

  1. BeforeAn AI with no access to the model can only advise from the outside.
  2. AI + BIMMCP exposes Revit operations, so the model can be queried directly for context.
  3. AutomationThe agent reads, modifies and creates through that same interface.
  4. VerifyIt checks its own work, and a person still signs off on the outcome.
Discuss this work

What if AI could understand a building?

Not a chatbot bolted onto a viewer. A system that reads the model as structured information, reasons about it, and proposes what to do next.

Conceptual building model

  • Reviewed
  • Attention
  • Issue

Analysis view

Example question

Find all spaces without air terminals.

42
spaces reviewed
8
require attention
3
potential issues

Spaces are read as enclosed volumes with properties, so "missing something" becomes a question the model can answer.

Conceptual AI analysis. The figures and the plan are illustrative demonstration data, not results from a real project.

The building becomes the interface.

AI across the construction workflow

This is not one feature in one application. The same idea applies at every stage where information moves.

  1. Design

    Options explored against real constraints.

  2. Engineering

    Sizing and analysis driven by model data.

  3. Documentation

    Drawings and schedules generated, not assembled.

  4. Coordination

    Gaps and clashes surfaced early and continuously.

  5. Construction

    Model information usable on site, not only in the office.

  6. Operations

    The asset stays queryable after handover.

AI Agents for Construction

AI that can reason about construction workflows and interact with the software that executes them.

Idle. Start the demonstration to step through the pipeline.

  1. Userstates the intent
  2. AI Agenttakes the request
  3. Understandreads the model context
  4. Reasonworks out what is required
  5. Plansequences the operations
  6. MCPthe interface to the tool
  7. Revitwhere the work happens
  8. Read / Modify / Createacts on the model
  9. Verifychecks the result

Interactive demonstration. This page is not connected to Revit — it steps through the concept, it does not execute anything.

The agent does not replace the authoring tool. It drives it, and then checks its own work.

The Human stays in control.

AI handles complexity.

Humans handle intent.

Construction carries real responsibility — for safety, for cost, and for what actually gets built. Automation should extend what a professional can do, not quietly take over the judgement they are accountable for. The system proposes. The engineer decides.

I build the systems behind this future.

  • BIM professional
  • Educator
  • Automation builder

I explore how AI can transform real-world construction workflows — from BIM automation and CAD integration to AI agents that can reason about the work itself.

Ideas are interesting. Working systems are better.

  • Revit AutomationTools that remove repeated modelling work.
  • CAD → BIMTurning existing drawings into model geometry and data.
  • AI × RevitIntent translated into model operations.
  • MCPGiving agents a real interface to authoring tools.
  • Gulshad BIM ToolkitA Revit add-in bringing the automation together.

Build. Learn. Explore.

The tools and the training come out of the work. They are how it gets shared, not the reason it exists.

Tools

Automation you can run today

Gulshad BIM Toolkit is the first of these: a Revit add-in built out of repeated project work, with more automation to follow.

Learn

Training on BIM and automation

Courses and training covering BIM workflows, Revit automation, and how to apply AI to construction work in practice.

Explore

Built in the open

Walkthroughs, tutorials and resources covering the same problems and the same workflows, shared as they are worked out.

Let's build the future of construction.

  • Buildings are becoming more connected and more data-rich.
  • BIM is becoming more than documentation.
  • AI is becoming more than a chatbot.
  • Automation is becoming part of everyday construction work.
Build togetherExplore the BIM ToolkitLearn