CASUS is a Swiss legal AI platform for law firms and in-house legal teams that analyses construction contracts and other documents directly in Microsoft Word or in the browser. The Risk & Quality Review identifies risks from each party's perspective, prioritises them by severity, and suggests concrete drafting options. All data stays in Switzerland or the EU, with no US transfer and no human review.
Why construction contracts are particularly error-prone
Construction contracts are among the most complex documents in private commercial law. A typical general contractor agreement covers detailed service descriptions, payment schedules, security deposits, warranty provisions, and acceptance conditions - all tied to technical standards, construction programmes, and project durations that can span several years.
Errors or ambiguities in these contracts have direct financial consequences. Disputed change orders, unclear termination clauses, or liability provisions without a cap can significantly inflate project costs. At the same time, legal teams and law firms are under pressure: contracts need to be reviewed quickly, often before a project officially starts.
The construction sector still lags behind other industries when it comes to digitisation. According to a study by ZAB Zukunftsagentur Bau GmbH and the University for Continuing Education Krems, the sector has enormous catch-up potential and many companies are only at the beginning of their AI adoption. That is precisely where AI-based contract analysis becomes useful.
Which risks appear most frequently in construction contracts
Construction contract risks can broadly be grouped into seven areas that should take priority in any first review:
Unclear service descriptions are the most common starting point for disputes. Phrases like "in accordance with generally accepted technical rules" sound precise but leave considerable room for interpretation.
Change order clauses without cost controls allow additional work without a defined price ceiling. Without a cost limit, this can lead to substantial overruns over the course of a project.
Acceptance criteria - when is a work deemed accepted? If a building is partially occupied before formal acceptance, this can unintentionally trigger warranty periods.
Warranty periods and their reduction are also critical. Under Swiss law, the standard period for structures is five years under OR Art. 371. Clauses that fall below this are high-risk.
Security deposits - a 10% retention is common in many contracts, but sits well above the market standard of around 5%.
Construction timeline clauses with vague wording like "approximately Q2" do not constitute a binding deadline, and therefore provide no basis for enforceable penalties.
Lump-sum vs. unit-price contracts are often not clearly distinguished, which causes ambiguity about compensation when quantities change or the scope shifts.
How AI-assisted contract review works for construction contracts
AI-assisted contract review identifies risky formulations in the document, assigns them to the relevant parties, and produces concrete improvement options - without losing legal meaning.
CASUS analyses construction contracts through the Risk & Quality Review: the system automatically identifies which parties appear in the contract and analyses risks from each party's perspective. Every finding is assigned a severity level (low / medium / high) and linked to a drafting option that can be applied directly in Word.
In practice: a security deposit clause without a clearly defined maximum is flagged as a risk, assigned to the relevant party, and matched with a drafting suggestion - all in one step, without switching to a separate research portal.
The Benchmark workflow adds another layer: a construction contract is checked against an internal playbook or established clause standards. This reveals whether a data protection clause is missing, whether liability is correctly capped, or whether a warranty exclusion meets the standard. The output includes a percentage match score.
When lawyers need to locate specific clauses or ask targeted questions within the contract, the AI Chat is available. Answers are linked directly to the relevant passages in the document - no more manually searching through a 60-page contract.
When there is uncertainty about the legal position on a specific risk, the Legal Research mode provides access to over 660,000 cantonal and federal court decisions. Results are structured and source-based, with direct inline previews of the relevant reasoning sections.
Law firms and legal departments that regularly review construction contracts can try CASUS for free. Sign up at app.getcasus.com/signup with no commitment required - directly in Microsoft Word or via the web app.
What changes in practice for legal departments
The time required for an initial contract review is one of the biggest practical challenges with construction contracts. An experienced lawyer can work through a 50-page works agreement in one to two hours - but that time is not available when several projects are starting simultaneously.
An AI-assisted first pass can significantly reduce this workload. The initial analysis flags the high-risk sections, prioritises them, and provides drafting options. The lawyer can then focus on the areas where professional judgement matters most - negotiation strategy, risk assessment, legal classification.
This is especially relevant for legal departments in construction companies or general contractors that regularly process large volumes of subcontractor agreements, supplier contracts, and GC contracts. For those cases, the AI Data Room is well suited: dozens of contracts can be uploaded in parallel, the relevant extraction fields are defined by prompt, and the result is a tabular clause matrix - ready to export.
A further benefit is consistency. When multiple people review contracts, different standards tend to emerge. A playbook-based benchmark ensures every contract is measured against the same reference.
Data protection and security for construction contract data
Construction contracts often contain commercially sensitive information: pricing models, project calculations, subcontractor structures. For companies operating in Switzerland, the question of where data is stored and processed is therefore relevant.
CASUS hosts all data in Switzerland or the EU. No data is transferred to the US. Zero Data Retention applies: uploaded documents are not stored after analysis. The system operates without human review - no employee on the provider's side reads the processed contracts.
Further details on the technical infrastructure and security certifications are documented on the CASUS security page.
CASUS for law firms with construction clients
Law firms that regularly handle mandates in the construction sector benefit from a standardised review methodology. The Benchmark workflow allows firms to store their own playbook for construction contracts. All contracts are then reviewed against this internal standard - with consistent findings and recommendations.
Lawyers advising clients during contract negotiations can use the improvement suggestions from the review directly as a basis for discussion. The drafting options are inserted into Word with correct formatting, ready to use.
After the main analysis, the Proofread workflow is worth running as a final step: it checks cross-references, definitions, numbering, and placeholders - typical error sources in long construction contracts that have gone through multiple revision rounds.
FAQ
Can AI fully review a construction contract from a legal perspective?
No. AI tools like CASUS produce structured risk analyses, clause comparisons, and drafting options - but not a conclusive legal opinion. Legal assessment and negotiation strategy remain the responsibility of qualified lawyers.
Which clause types in construction contracts does AI identify?
CASUS identifies, among others, liability clauses, security deposits, acceptance provisions, change order rules, warranty periods, and contractual penalties. The depth of review depends on the fields and parameters defined in the review or benchmark.
How long does an AI-assisted first review of a construction contract take?
The initial CASUS analysis typically takes a few minutes, depending on the length of the document. The output includes prioritised findings with severity ratings and drafting suggestions.
Is CASUS designed for Swiss construction law?
CASUS is oriented toward the Swiss legal context: Swiss hosting, no US data transfer, and spelling conventions aligned with Swiss German standards. Legal research queries can draw on over 660,000 Swiss cantonal and federal court decisions.
Can CASUS review multiple subcontractor agreements at the same time?
Yes. With the AI Data Room, many contracts can be uploaded in parallel and analysed according to self-defined fields. The output is a structured table - suitable for compliance checks, due diligence, or managing large contract portfolios.
What data standards apply when using CASUS?
CASUS processes data exclusively in Switzerland or the EU, does not transfer data to the US, and does not retain uploaded documents after analysis. Zero Data Retention and no human review by provider staff apply.
Can CASUS be integrated into an existing contract review process?
CASUS runs directly in Microsoft Word as an add-in and through a web app. An internal playbook can be stored as a reference standard. No separate system integration is required.
Which standards and norms are relevant for construction contracts in Switzerland?
For private-law construction contracts in Switzerland, the primary references are the OR (in particular Art. 363 ff. on contracts for work) and the SIA standards (e.g. SIA 118). Companies with international projects also use FIDIC contracts.
