# AEC AI Workflow Starter Kit

## 7 practical workflows for architects, engineers, BIM/VDC teams, project managers, and AEC firm leaders.

Version 1.0 | 2026  
VexASI Academy

VexASI Academy provides practical AI operations training for AEC professionals and firms. This starter kit is a first-pass operating resource for teams that want safer AI usage, clearer review habits, and workflows that can be tested without pretending AI outputs are automatically correct.

Intended users:

- Architects and design teams
- Engineers and technical discipline leads
- BIM/VDC teams
- Project managers
- AEC firm leaders
- Business development and marketing teams inside AEC firms

How to use this kit:

1. Pick one workflow with low confidentiality risk.
2. Define the source material, owner, expected output, and review step.
3. Run a controlled test with non-sensitive or approved internal material.
4. Document what worked, what failed, and what must be reviewed by a qualified person.
5. Decide whether the workflow is ready for a broader pilot, needs revision, or should be retired.

## 1. Safe AI Usage Rules for AEC Teams

AI tools can help with drafting, summarizing, classification, comparison, and workflow support. They should not be treated as independent professional judgment.

Core rules:

- Do not paste confidential client, project, employee, financial, or proprietary data into unapproved tools.
- Do not use AI output as final project advice without qualified human review.
- Do not rely on AI for code compliance, life safety, sealed documents, or project-specific professional conclusions.
- Keep original source material attached to any AI-assisted output that affects project records.
- Label AI-assisted drafts clearly during internal review.
- Confirm facts, dates, names, standards, quantities, and references against source material.
- Use firm-approved tools and follow internal data-handling rules.
- Escalate unclear, high-risk, or client-sensitive use cases before testing.

Minimum review gate:

- Source material is known and approved for the tool.
- Output has an assigned human reviewer.
- Reviewer checks facts, assumptions, omissions, and tone.
- Final use is documented according to firm policy.

## 2. Meeting Notes to Action Log Workflow

Purpose: Turn project meeting notes into a structured action log that can be reviewed by the project manager before distribution.

Best-fit inputs:

- Internal meeting notes
- Approved transcript excerpts
- Agenda and attendance list
- Prior action log

AI-assisted steps:

1. Ask the tool to extract decisions, open questions, action items, owners, due dates, and dependencies.
2. Ask it to flag unclear ownership, missing dates, and unresolved decisions.
3. Ask it to compare the new notes against the prior action log.
4. Have the PM review every action item before it enters the project record.

Suggested output fields:

- Action item
- Owner
- Due date
- Source note
- Status
- Dependency
- Risk or escalation needed

Review checklist:

- Are owners named correctly?
- Are dates present and accurate?
- Are decisions separated from discussion notes?
- Are unresolved items clearly marked?
- Does the output preserve project context without inventing facts?

## 3. RFI/Submittal Drafting Workflow

Purpose: Support first-draft structure for RFIs, submittal comments, and response summaries while keeping qualified review in control.

Best-fit inputs:

- Approved RFI or submittal text
- Relevant spec excerpts
- Drawing references supplied by the project team
- Prior approved response examples

AI-assisted steps:

1. Ask the tool to summarize the request, affected scope, cited references, and missing information.
2. Ask it to draft a neutral response structure or reviewer questions.
3. Ask it to identify assumptions and references that need confirmation.
4. Route the draft to the responsible project professional for review and revision.

Review checklist:

- Are all references traceable to supplied material?
- Did the tool invent requirements, dimensions, products, dates, or approvals?
- Is the response framed as a draft for review?
- Does the responsible professional agree with the conclusion?
- Is the final record stored according to firm and project policy?

Do not use this workflow to bypass professional review, contract requirements, client review, or authority having jurisdiction processes.

## 4. Drawing Review / QA Checklist Workflow

Purpose: Help project teams organize review comments and checklist coverage before formal QA/QC review.

Best-fit inputs:

- Firm QA checklist
- Sheet list
- Drawing issue log
- Discipline coordination notes
- Public or approved internal standards

AI-assisted steps:

1. Ask the tool to convert a QA checklist into a discipline-specific review table.
2. Ask it to map comments by sheet, discipline, severity, owner, and status.
3. Ask it to identify missing fields, duplicate comments, and vague issue language.
4. Have discipline leads review and approve the final comments.

Suggested output fields:

- Sheet or model area
- Discipline
- Issue summary
- Source comment
- Severity
- Owner
- Status
- Required follow-up

Review checklist:

- Does each comment point to a real source?
- Are severity labels consistent with firm practice?
- Are comments actionable?
- Are assumptions separated from confirmed issues?
- Has a qualified reviewer checked technical implications?

## 5. BIM Data Validation Workflow

Purpose: Support structured review of model data exports, naming consistency, and coordination dashboards.

Best-fit inputs:

- Approved model data exports
- Equipment schedules
- Room data sheets
- Naming standards
- Clash or coordination logs

AI-assisted steps:

1. Provide a structured export, such as CSV, along with the validation rule set.
2. Ask the tool to identify missing values, inconsistent naming, duplicate IDs, and outlier fields.
3. Ask it to group exceptions by discipline, system, level, zone, or responsible team.
4. Route the exception list to the BIM/VDC owner for validation.

Review checklist:

- Are validation rules explicitly supplied?
- Did the tool preserve IDs and source rows?
- Are exceptions grouped in a useful way?
- Are false positives documented?
- Are downstream model or dashboard changes made only through approved workflows?

## 6. Proposal and Business Development Workflow

Purpose: Help marketing and business development teams assemble pursuit research, win themes, and draft outlines with clear review.

Best-fit inputs:

- Public client information
- Approved firm experience
- RFP or RFQ text
- Interview notes approved for internal use
- Existing proposal boilerplate

AI-assisted steps:

1. Ask the tool to summarize client priorities and stated evaluation criteria from source material.
2. Ask it to map relevant firm experience to those priorities.
3. Ask it to draft a proposal outline, interview prep questions, or pursuit briefing.
4. Have BD, marketing, and technical leaders review claims before external use.

Review checklist:

- Are all client claims source-backed?
- Are firm experience claims accurate and approved?
- Does the draft avoid unsupported guarantees?
- Are differentiators specific to the pursuit?
- Has the technical team reviewed scope-sensitive language?

## 7. Firm AI Readiness Checklist

Purpose: Help leadership decide whether the firm is ready to expand AI usage beyond isolated experiments.

Readiness areas:

- Policy: The firm has clear rules for approved tools, restricted data, review expectations, and escalation.
- Ownership: AI adoption has named business, technical, and risk owners.
- Training: Role groups know what AI can and cannot be used for.
- Workflow selection: Pilot workflows are selected by value, repeatability, and risk level.
- Review gates: Consequential outputs require qualified human review.
- Data boundaries: Teams understand client, project, employee, financial, and proprietary data restrictions.
- Tool control: The firm knows which tools are allowed and why.
- Measurement: Pilots track time saved, quality issues, adoption, rework, and failure cases.
- Documentation: Prompt patterns, workflow steps, and review notes are captured for reuse.
- Roadmap: The firm has a practical sequence for pilot, revise, scale, or retire decisions.

Scoring prompt:

- Green: Clear policy, owner, approved tools, and review path exist.
- Yellow: Some pieces exist, but teams are inconsistent or informal.
- Red: Usage is unmanaged, unclear, or dependent on individual experimentation.

## Disclaimer

This starter kit is educational material for professional training and workflow planning. It is not legal advice, code-compliance advice, sealed architectural or engineering review, project-specific professional advice, or a substitute for qualified professional judgment. AI outputs can be incomplete, inaccurate, or misleading and must be reviewed before use in project, client, contractual, regulatory, or business-critical contexts.

VexASI Academy is designed with professional training standards in mind and may support a future CE-ready pathway. This resource does not claim AIA CE approval, HSW approval, state licensure credit, certification status, or approved continuing education.

For firm training, launch updates, or role-specific AI workflow support, contact VexASI Academy: vexasi.com/academy or intel@vexasi.com.