Artificial intelligence (“AI”) is no longer a technology that businesses experiment with. Globally, AI adoption is accelerating, and recent data suggest that most companies are either already using AI or actively planning its implementation, with many reporting measurable gains in productivity and revenue. This shift is also reshaping how mergers and acquisitions (“M&A”) are evaluated, negotiated and concluded in South Africa. As AI becomes more deeply embedded in business operations, M&A transactions must evolve to treat AI not only as a strategic asset but also as a practical tool that influences the transaction process itself. 

AI as part of the deal

When a target company develops, owns or depends on AI systems, understanding those systems is essential before finalising a transaction. AI that materially contributes to enterprise value cannot be reviewed as though it were a routine technology asset. Its governance structures, performance reliability and associated risks directly affect valuation and the outcome of due diligence.

International frameworks provide helpful benchmarks for assessing these risks. The EU High-Level Expert Group on AI identifies lawfulness, ethical alignment and technical robustness as core requirements for trustworthy AI. The NIST AI Risk Management Framework complements this by categorising risks, such as technical failures, human misuse and broader systemic impacts, and promoting structured processes to identify, measure and mitigate them.

While these frameworks are not binding in South Africa, they are increasingly relied upon in cross-border transactions and offer practical guidance for evaluating AI-related exposures.

Intellectual property and ownership considerations

AI-driven businesses derive significant value from algorithms, training data and system outputs. Determining ownership and control of these elements is central to valuation and post-merger integration planning.

During due diligence, particular attention should be paid to:

• ownership of source code and algorithms
• assignment of intellectual property created by employees and contractors
• confidentiality, non-disclosure and restraint-of-trade provisions protecting proprietary systems
• licensing arrangements for third-party AI tools
• the legal status and protectability of AI-generated outputs

Because many commercial contracts predate widespread AI adoption, they often fail to address ownership of AI-generated content or derivative works. This can create uncertainty that materially affects valuation and can complicate integration after closing.

AI in the transaction process

AI is not only a due diligence subject. It is also changing how due diligence itself is performed. Tools used for document review, clause extraction and large dataset analysis are increasingly AI-assisted, offering significant efficiency gains and aligning with broader corporate trends toward AI-enabled operational optimisation.

However, these tools still require human oversight. Continuous evaluation, verification and accountability remain essential, especially in high-stakes environments such as M&A transactions. Professional judgment and structured governance are not optional; they are necessary safeguards.

Generative AI systems may produce inaccurate or fabricated information - often referred to as “hallucinations.” In an M&A context, reliance on incorrect outputs could result in missed risks or flawed assessments with severe financial consequences. To mitigate this, practitioners should consider:

• manually sampling AI-reviewed documents
• testing outputs across multiple contract types and risk categories
• prioritising high-value or high-risk agreements for enhanced review

The way forward for South African M&A

AI is transforming M&A transactions in two fundamental ways.

Firstly, it has become a complex and highly valuable asset class, bringing with it specialised legal, regulatory and valuation considerations that require careful, structured assessment.

Secondly, AI is reshaping transaction execution, offering efficiency gains while introducing new risks relating to data accuracy, confidentiality and oversight.

As global standards evolve and are driven by bodies such as the EU High-Level Expert Group on AI and the U.S. National Institute of Standards and Technology, South African practitioners must remain aligned with emerging international best practice.

To succeed in this environment, M&A advisors must develop dual capability: the ability to evaluate AI-driven businesses as acquisition targets and the competence to responsibly deploy AI tools when managing the transaction process. In a competitive market increasingly shaped by technological capability, AI is no longer peripheral to M&A -it is now central to modern M&A strategy.