A Share Purchase Agreement (SPA) is the central contract in any M&A transaction. It does not simply confirm a sale – it distributes risk, defines warranties, and sets out everything that applies between signing and closing. Large SPAs run to several hundred pages, and every one of those pages needs careful legal analysis. That is exactly where AI-assisted contract review comes in.
This article explains how AI can support SPA review and M&A due diligence, which clauses tend to be most error-prone, and what law firms and in-house teams can practically gain.
What an SPA covers – and why reviewing one takes so long
The SPA translates the commercial logic of a deal into enforceable contractual mechanisms. It typically covers:
Purchase price and adjustment mechanisms (closing accounts, locked box, earn-out)
Representations and warranties – the seller's assurances about the state of the business
Indemnities – liability provisions for specific, known risks
Covenants – behavioural obligations between signing and closing
Closing conditions – preconditions that must be met before the transaction completes
Non-compete clauses and governance arrangements
Every one of these areas has negotiation potential. And every one of them, if drafted ambiguously or incompletely, can lead to disputes later – for instance when the parties develop different interpretations of the closing accounts mechanism.
There is another layer of complexity in share deals specifically: the buyer acquires not just the assets but also all hidden liabilities. That makes thorough SPA review even more pressing than in an asset deal, where liabilities can be selectively excluded.
Typical weaknesses in an SPA
M&A practice reveals a number of recurring problem patterns.
Liability caps are missing or unclear. When warranties are agreed without a cap, or the cap is worded in a way that requires interpretation, significant legal uncertainty follows.
Accounting principles are insufficiently defined. In a closing accounts mechanism, everything depends on which accounting standards apply. Without a clear definition, the parties end up disputing the working capital calculation after closing.
Definitional inconsistency. In long SPAs produced through multiple drafting rounds, the same term sometimes appears with slightly different wording in different places. This is a classic drafting problem.
Missing or incomplete clauses. Data protection, IP ownership, termination rights in key contracts – these topics can fall through the cracks in the pressure of a fast-moving transaction.
How AI changes SPA review
AI-assisted contract analysis can substantially reduce the time spent on manual review. CASUS, a Swiss legal AI platform, offers several complementary modules for this purpose.
Risk analysis at clause level
The Risk & Quality Review analyses a contract from the perspective of a defined party. The output is a structured list of findings, each with a severity rating (low / medium / high), assignment to the affected party, and concrete drafting suggestions. Those suggestions can be applied directly in Word without copy-paste.
For SPA review, this means: instead of reading the document linearly, the team receives a prioritised overview of the most critical clauses. Liability gaps, ambiguous warranty language, or missing covenants surface early.
Comparing against a playbook
The Benchmark workflow checks an SPA against a defined reference standard – an internal firm playbook or a best-practice template. CASUS shows which standard clauses are missing, which are insufficiently developed, and where the document deviates from the standard. A percentage match score gives an immediate sense of how far the draft SPA sits from the reference.
This is particularly useful for SPAs: topics such as locked-box leakage protection, tax clauses, or IP warranties can be checked systematically, without anyone having to work through a checklist manually.
Parallel document analysis in the data room
M&A transactions rarely involve just one document. The due diligence process often brings dozens or hundreds of contracts – leases, supplier agreements, employment contracts, licence agreements – all requiring review.
The AI Data Room supports the upload of large document volumes. The extraction fields are defined by the team via prompt – for example "liability caps", "notice periods", or "data protection clauses". The result is a tabular overview that can be exported directly to Excel. Anomalies – such as liability clauses without a cap or notice periods exceeding twelve months – are flagged and prioritised by risk.
Legal research directly in the workflow
When a legal question arises during SPA review – say, the enforceability of a non-compete under Swiss law or the tax treatment of an earn-out – the Legal Research mode within AI Chat can help. It searches across more than 660,000 cantonal and federal court decisions as well as statutory provisions, and delivers structured, source-based assessments that feed directly into work product.
Results are linked to the relevant passages; users can jump to the source without leaving the platform.
Consistency and formal quality
Before an SPA goes into negotiation or to the counterparty, a final formal check is worthwhile. The Proofread workflow checks cross-references, definitions, annexes, and placeholders, flags inconsistent terminology, and verifies that numbering and references are correct. This is not a substitute for legal review, but it is effective protection against embarrassing errors in the final document.
What this means for law firms and in-house teams
Time pressure in M&A transactions is a given. Delivering a first risk assessment of a 200-page SPA in hours rather than days requires tools that do more than keyword search – they need to analyse with structure.
AI does not replace legal expertise here. What it changes is the proportion of time spent on mechanical review versus substantive legal judgment. A team that already has a structured overview of the critical clauses can go deeper on the issues that actually matter.
For law firms, there is also a commercial dimension: teams that work more efficiently can handle a larger number of mandates in the same time without reducing quality.
CASUS is hosted in Switzerland and the EU, with no data transfer to the US. The platform operates without human review and without data retention after a session – both are requirements that commonly arise in M&A processes involving sensitive business information. Full details are available on the security page.
Try it on a real document
CASUS can be used directly in the browser or as a Microsoft Word add-in. Law firms and in-house teams can start for free – no sales call required, no implementation project needed. Upload a concrete SPA or due diligence document and see what the analysis produces: Start for free.
FAQ
What is a Share Purchase Agreement (SPA)?
A Share Purchase Agreement (SPA) is the central contract in M&A transactions, in which buyer and seller establish all material terms of the business acquisition. It covers purchase price, warranties, liability, closing conditions, and behavioural obligations through to completion.
How does a share deal differ from an asset deal?
In a share deal, the buyer acquires shares in the company and with them all assets and liabilities – including hidden ones. In an asset deal, specific assets are transferred individually; unknown liabilities remain with the seller, but the administrative complexity is higher.
Which clauses in an SPA are particularly critical?
The most error-prone areas are liability caps, the definition of accounting principles in closing accounts mechanisms, IP warranties, tax clauses, and the drafting of non-compete provisions. Gaps or ambiguities in these areas frequently lead to post-closing disputes.
How can AI help with SPA review?
AI can analyse an SPA systematically for risks, compare clauses against a reference standard or playbook, identify missing topics, and handle consistency checks. This reduces the time spent on manual review and gives lawyers a structured overview earlier in the process.
What does the AI Data Room do in M&A due diligence?
The AI Data Room supports parallel analysis of many documents. Teams define which information to extract – for example liability clauses, notice periods, or data protection provisions. The output is a tabular summary across all documents, with anomalies flagged and prioritised by risk.
Does AI replace the legal review of an SPA?
No. AI analysis supports legal work but does not replace legal judgment. What changes is the pace of the mechanical groundwork, leaving more time for substantive legal assessment.
How secure is CASUS when working with sensitive M&A documents?
CASUS is hosted in Switzerland and the EU. There is no data transfer to the US, no human review of documents, and no data retention after the session. These features align with what M&A processes typically require when handling confidential business information.
Can CASUS research the legal position on SPA clauses?
Yes. The Legal Research mode within AI Chat searches across more than 660,000 cantonal and federal court decisions as well as statutory provisions. Results are source-based and structured – they can be used directly to support clause rationale or for internal assessments.
