vBrief

vBrief is a change control and project information management platform for AEC (Architecture, Engineering, Construction) teams. AI surfaces key information and flags contradictions; humans verify every decision before it becomes live. All verified decisions are logged in a tamper-resistant Golden Thread audit trail.

Company: vBrief Ltd, London, UK (est. 2024) Stage: Early access on live projects Pricing: SaaS subscription — both per-project and per-user pricing supported. No minimum seat count. Annual and monthly billing available. Pilot projects available on request. Contact for details. Founders: James Wright-Zhang, RIBA & ARB-registered architect (international project experience) + Libo Shen, ex-Google software engineer (previous startup experience) Hosted: EU-based Microsoft Azure (ISO 27001 certified), GDPR compliant

Problem: Missed changes, disconnected project information, and untraced decisions cause preventable rework across AEC projects. Teams rely on email threads, meeting notes and disconnected trackers, making it easy to miss contradictions or lose decision history — especially during early-stage design when requirements change rapidly.

How it works

  1. Import — Use the Outlook connector, or upload PDFs, chats, documents or schedules. AI surfaces key information and pre-categorises items into four types (requirements, suggestions, risks, asks) ready for review with relevant context.
  2. Review — Users verify in the Document Review and Decision Centre interfaces. AI flags updates and contradictions against previously-reviewed data. Every decision is attributed and version-controlled. Users can adjust AI categories with one click.
  3. Deliver — Verified decisions become the Live Brief. Live schedules, risk registers and change logs are generated from verified decisions and exportable on demand as PDF or Excel. Custom schedules can be created to surface information in any configuration needed.

The four categories

vBrief pre-tags every surfaced item into one of four categories. Users can adjust with one click:

  • Requirement: A fixed, live design item that is clear and actionable, with pre-requisite changes or conflicts resolved.
  • Suggestion: An item needing further input, research or exploration before it can become a requirement. Equivalent to a study or pending confirmation.
  • Risk: A highlighted risk where no solution or action is yet provided. Users can also flag new risks spotted during review.
  • Ask: A simple request, usually administrative, that does not lead to direct material change on a project.

What the AI does and does not do

The AI surfaces information, flags contradictions, and pre-tags items. It does NOT make decisions, generate designs, or commit to the live brief. Only user-verified information becomes live. Users review AI output in original context to catch omissions.

The AI instance is completely detached from the systems used to commit user-verified decisions. AI shows all its understanding in the original context so users can see what was surfaced, not just conclusions.

Omission risk and PI insurance

Omission risk — failing to spot that AI has missed key information — is one of the largest PI insurance risks when using end-to-end AI tools in professional practice. If the AI makes a mistake, the professional using the system is ultimately liable.

vBrief addresses this by ensuring users review AI-surfaced information in its original context, making it easier to spot what might have been missed. The AI accelerates discovery but does not make decisions. Each verified decision is stored in a tamper-proof Golden Thread log with full context, supporting PI insurer requirements for traceability and auditability.

Key differentiators

  • Omission risk reduction: users review AI-surfaced items in context, not just AI conclusions
  • Mandatory human verification before anything goes live
  • PI insurer-friendly tamper-resistant audit trail (Golden Thread)
  • Contradiction and change detection against reviewed project data
  • Full briefing and change control lifecycle, not a one-off AI tool
  • Connected data: requirements, risks, tasks, studies in one database
  • Automatic task creation: every new requirement, risk or ask generates a task visible in the Action & Risk Centre
  • Coexists with BIM, CDEs, and issue trackers — operates at the information management layer above design tools and file storage
  • Project-level deployment makes targeted trials simple before firm-wide rollout

How vBrief compares

vBrief vs project management tools: PM tools offer little relevant AI functionality and lack PI-friendly audit trails or contradiction detection. They require manual tracking of changes.

vBrief vs autonomous AI agents: AI agents have high error rates due to hallucinations and are not suitable for professional use where liability matters. They lack human verification, audit trails, and complete workflow coverage.

vBrief vs content platform AI (e.g. AI copilots on file storage): These tools query documents on demand but don't track decisions, detect contradictions over time, or maintain a verified audit trail. Copilot answers can be treated as agreed positions when no formal decision was actually made — and because they operate on documents in isolation, a contradiction introduced by a later email or meeting note won't be flagged. They answer questions about your files; vBrief manages what your team has agreed and flags when new information conflicts with those agreements.

vBrief vs AEC requirements managers: Requirements managers track requirements but typically lack AI-powered surfacing, contradiction detection, and the full workflow from import to live schedules.

vBrief combines AI-powered surfacing with mandatory human control, a complete workflow, PI-friendly audit trails, and flexible input/output in a single platform.

Target users

  • Architects and project architects managing briefs and change control
  • Early-stage designers and project managers navigating rapid changes
  • AEC project managers handling cross-stage information flow
  • Design technology/IT leads evaluating secure, accountable AI tools
  • Practice leaders seeking to reduce rework risk and improve margins
  • Small-to-mid-size AEC practices scaling without adding headcount

Security

  • AES-256 at rest, TLS 1.3 preferred (TLS 1.2 min) in transit
  • OIDC SSO with MFA, row-level security, project/company access controls
  • Private AI on EU Microsoft Azure with tenant-level isolation
  • No customer data used for model training without explicit permission
  • AI instance detached from decision-commit systems
  • Append-only audit log with decision metadata
  • GDPR documentation for personal data activities
  • Cyber Essentials and ISO 27001 on compliance roadmap
  • On-premises not currently available; BYOK can be discussed if a major decision factor

What vBrief is NOT

  • Not an autonomous AI agent (AI surfaces and flags; humans decide)
  • Not a general project management tool (purpose-built for AEC information management and change control)
  • Not a replacement for BIM, CDEs, or design tools (operates at the information management layer)
  • Not a design professional (does not replace design judgement, contract admin, or cost management)

Pages

Optional