Jin Lai / Architectural designer

Architecture.
BIM. & AI.

An architectural foundation.
An evolving practice.

Explore possible AI-assisted solutions for BIM production and construction documents, alongside my M.Arch academic background.

Physical model of Adaptive Climate Machine, Jin Lai's HKU M.Arch thesisM.Arch thesis / 2024
Adaptive Climate MachineFrom the academic archive ↗
AI-assisted solutions

BIM production.
Construction document support.

Six possible project scopes for architectural and construction teams. Select a solution to explore the inputs, process and review needed.

Choose another solution ↑
BIM productionPossible project scope

BIM model auditing and data cleanup

Find information gaps before they become coordination problems.

Inputs

A Revit or IFC model, agreed information requirements, naming rules and the project stage to be checked.

Workflow

  1. Agree the checks

    Translate the information requirements into explicit rules for required parameters, identifiers, naming and record completeness.

  2. Inspect model data

    Extract element data with BIM tools or scripts. Run the agreed checks and link each finding to its model element.

  3. Review corrections

    Group the issues, propose corrections and review them with the model owner before any approved changes are applied.

AI contribution

AI can assist in interpreting requirements, developing checking scripts and explaining findings. The checks execute against extracted model data through explicit rules.

Deliverables

  • Element-linked issue report
  • Missing-value, duplicate-ID and naming check tables
  • Reviewed correction plan and unresolved-issue list

Review method

Test the rules on known examples, trace findings back to elements and spot-check the extracted data. The project team agrees corrections and confirms acceptance; an information audit does not establish overall model or design compliance.

Back to solutions ↑
Choose another solution ↑
BIM productionPossible project scope

Revit production automation

Turn repeatable drawing tasks into controlled production steps.

Inputs

A Revit model, sheet register, office standards, target export settings and an agreed list of tasks and parameters that may be changed.

Workflow

  1. Define the task

    Confirm sheet creation, naming, selected parameter updates or export requirements, including exceptions and permitted changes.

  2. Build and test

    Develop automation using Revit's supported interfaces. Test on a model copy and compare a small output set with the agreed standards.

  3. Run and record

    Execute approved tasks, record completed actions and exceptions, and check the resulting sheets and exports.

AI contribution

AI assists tool development, code explanation and test planning. The tested tools execute production actions through Revit's supported interfaces; AI text alone does not modify or validate the model.

Deliverables

  • Reusable production tools and usage notes
  • Generated sheets or drawing export package
  • Task and exception log

Review method

Validate on a model copy, inspect representative sheets and parameter changes, and reconcile the outputs with the sheet register. The project team reviews the drawing package before issue. Tool compatibility and scope depend on the Revit version and model setup.

Back to solutions ↑
Choose another solution ↑
BIM productionPossible project scope

Facade and panel production workflows

Connect facade geometry to usable panel information.

Inputs

A suitable facade model with identifiable panel elements, material information, location references and agreed measurement and numbering conventions.

Workflow

  1. Check the model setup

    Confirm which elements represent panels, how they are located and which dimensions and quantities the model can support.

  2. Extract and calculate

    Use BIM/CAD tools or explicit scripts to derive panel IDs, dimensions, areas, materials and locations. Flag geometry or data that cannot be resolved.

  3. Organize production information

    Compile registers, quantity schedules and identification diagrams with traceable links to panel elements and exceptions.

AI contribution

AI can assist script development, register organization and explanation of exceptions. Geometry, measurements and quantities come from the model or explicit calculations.

Deliverables

  • Panel register with IDs, dimensions, areas, materials and locations
  • Quantity schedules and panel identification diagrams
  • Geometry and information exception report

Review method

Reconcile panel counts and IDs, sample dimensions against the model and verify units, area conventions and material mappings. The project team reviews the outputs. Fabrication suitability requires the responsible specialist's review of tolerances, joints and detailing.

Back to solutions ↑
Choose another solution ↑
Construction document supportPossible project scope

Drawing revision review

Make changes between issued drawing versions easier to follow.

Inputs

Two issued drawing versions, their sheet registers and revision information, with matching sheets and legible source files.

Workflow

  1. Match the versions

    Confirm issue dates, pair corresponding sheets and align the drawings so scale or position changes do not obscure the comparison.

  2. Compare and categorize

    Use drawing comparison tools to detect differences. AI assists in organizing the detected changes by location, type and possible coordination relevance.

  3. Prepare review notes

    Link annotated comparisons to a change register and identify potentially affected drawings and follow-up questions.

AI contribution

AI helps categorize detected differences and draft concise summaries. Each summary stays connected to the compared sheets; unclear changes remain questions for review.

Deliverables

  • Annotated drawing comparisons
  • Source-linked change register
  • Potentially affected drawing list and follow-up questions

Review method

Check the annotations against both issued versions and confirm summaries with the responsible designer. Drawing quality and alignment affect detection. A visual difference alone does not establish design intent or a contractual change.

Back to solutions ↑
Choose another solution ↑
Construction document supportPossible project scope

Specification and supplier-submittal pre-review

Put requirements and supplier proposals side by side.

Inputs

Relevant specification clauses and project requirements, supplier data sheets, certificates and submittal documents, with their issue versions.

Workflow

  1. Locate the requirements

    Identify the clauses and criteria relevant to the proposed product, recording document names, versions and page or clause references.

  2. Compare the proposal

    Use AI-assisted extraction and comparison to map supplier information to each requirement. Identify possible mismatches and missing evidence.

  3. Prepare a review table

    Present requirement-versus-proposal entries with both source references and specific requests for missing information.

AI contribution

AI assists document extraction, source-referenced comparison and draft review notes. Unsupported or ambiguous statements are flagged for review instead of treated as confirmed compliance.

Deliverables

  • Requirement-versus-proposal comparison table
  • Potential mismatch and missing-information register
  • Source references and draft clarification requests

Review method

Verify every comparison against the cited source and check versions, units and product variants. The responsible designer determines technical acceptance. Pre-review depends on supplied evidence and does not replace specialist assessment or product approval.

Back to solutions ↑
Choose another solution ↑
Construction document supportPossible project scope

Project document search and RFI preparation

Turn scattered project records into referenced questions.

Inputs

Issued drawings, specifications, meeting records and project queries, with document versions and an agreed set of records to search.

Workflow

  1. Organize the records

    Index the supplied documents with issue dates and references, and identify superseded or missing records.

  2. Retrieve relevant evidence

    Search for information relating to the query. AI helps connect retrieved passages and identify conflicting or incomplete instructions.

  3. Draft for coordination

    Prepare a referenced draft RFI or coordination summary, separating source evidence from questions and proposed follow-up.

AI contribution

AI assists retrieval, referenced drafting and issue organization. Answers are limited to the supplied records; missing evidence is recorded explicitly.

Deliverables

  • Draft RFIs with drawing, page or clause references
  • Referenced coordination summaries
  • Issue register with open questions and proposed follow-up

Review method

Open the cited sources, confirm the issue versions and review whether the draft accurately states the question. The project team approves and issues every communication. Document search does not resolve conflicting instructions or authorize design changes.

Back to solutions ↑

Scopes are agreed around the available model or documents and the project team's requirements. Technical acceptance and issuing communications remain with the project team.

Discuss a scope
Contact

Let's start a conversation.

Have a project in mind, or an idea to explore? Get in touch about design, BIM, or AI workflow collaboration.

Jinlaimail@gmail.com