As the fourth pillar of SRHD Concierge, Lighthouse Risk Check brings a claim-informed risk lens that general AI tools simply don’t have.
The four pillars that differentiate Spec Rep Help Desk
Spec Rep Help Desk Concierge is built on four pillars. Each one is designed specifically for Australian architects, and each one is a real differentiator versus general AI tools.
1) Building code information, Australian Jurisdictions only
Concierge keeps building code guidance anchored to the Australian context. It prompts for jurisdiction when it matters, references the NCC by edition where applicable, and avoids generic overseas assumptions. The goal is not to “certify” compliance, but to help you navigate the right pathway, ask the right questions, and document decisions clearly.
2) Building product information, built for specification
Concierge is designed to return the product evidence architects actually need, fast. Instead of vague marketing claims, it focuses on the practical inputs that de-risk specification: the right documents, the right performance checks, and the right follow-up questions for your approvals pathway. It is a specification support tool, not a sales chatbot.
3) Sustainability information, made more accurate through reputable eco labels
Concierge’s sustainability layer is strengthened through partnerships with reputable eco labels to improve the accuracy and completeness of environmental data. This helps architects compare options with more confidence, request the right evidence early, and avoid sustainability decisions being reduced to incomplete or inconsistent information.
4) Risk information, powered by Lighthouse Risk Check
Most AI tools are built to sound confident. That’s useful for drafting and brainstorming, but it’s not what architects need when decisions will be built, inspected, and potentially disputed years later.
Lighthouse Risk Check is SRHD Concierge’s risk engine. It is designed to do one job well: help architects see the common failure pathways early, and translate them into documentation and verification steps the project team can actually use. It’s a risk flagging and verification assistant that works alongside the Concierge’s code, product, and sustainability outputs.
What Lighthouse is based on
Lighthouse is not a generic “best practice checklist”. It has been shaped by three inputs that general AI tools don’t have:
Professional indemnity claim drivers: We’ve incorporated input from professional indemnity insurers and brokers about the recurring issue-types most likely to lead to disputes or claims in architectural practice.
Published defect guidance: We draw on publicly available defect research and guidance, including resources such as the NSW Government’s Building Defects Library published by Building Commission NSW, which consolidates common defect types and where they typically arise in documentation and construction.
Public dispute and litigation patterns: We’ve reviewed publicly available dispute narratives and decisions to understand how problems actually emerge on projects — interface failures, missing details, late substitutions, unclear responsibilities — not just how they are meant to work on paper.
This is not legal advice and it is not a compliance determination engine. It is a structured way to surface risk early, when it is still cheap to manage.
What it does in practice
Lighthouse runs quietly in the background while you use Concierge normally.
When your query touches a higher-risk area, Concierge still answers your question first. Then Lighthouse adds a compact footer that flags what matters most at a drawing-blind level:
Why higher risk: the likely failure mechanism (water path, vapour path, movement, fire/smoke path, geometry/clearances, coordination failure)
Document: the minimum details and spec notes that prevent site assumptions
Verify: one or two hold points before work is concealed or decisions become costly to reverse
If you want more detail, you can expand it on request. The expanded output provides a structured checklist, targeted confirmation questions, and user-facing external sources where available (NCC/ABCB/regulators/manufacturer manuals). Internal training documents are never shown as sources.
It catches the adjacent risks that general AI misses
Architects rarely ask “What might go wrong?” They ask focused questions like:
“What lining should I use?”
“Can I make this shower hobless?”
“What’s a good cladding option?”
General AI tools answer the narrow question and stop. Lighthouse is designed to also catch the adjacent claim drivers that often sit next to the question:
A sports hall question about linings can still hide a court clearance or set-out risk.
A window question can hide water path and condensation risk.
A substitution question can hide evidence and approvals pathway risk.
A demolition question can hide scope boundary and existing conditions risk.
This is where Lighthouse earns its keep: it helps you avoid the “looks fine until it fails” category of problems.
What you get as an architect
Lighthouse is built to support how architects actually work:
Faster documentation: prompts the minimum details and schedules that stop site assumptions.
Better coordination: highlights interface responsibilities and sequencing risks early.
Cleaner approvals pathway: nudges evidence collection before it becomes a scramble.
More defensible process: encourages decision records, hold points, and verification before concealment.
What Lighthouse is not
To keep it safe and useful, Lighthouse has hard limits:
It does not certify compliance.
It does not replace the project team, the certifier, or manufacturer installation instructions.
It does not provide legal advice.
It does not pretend to have seen your drawings.
It’s a practical risk flagging layer built into SRHD Concierge — one of four pillars that make the platform purpose-built for Australian architectural practice.