Artificial intelligence is rewriting the rules of mergers and acquisitions, forcing dealmakers to reconsider what makes businesses valuable. Javi Rodriguez, Global Head of Strategy at KPMG International, explains why AI defensibility is becoming a critical factor in assessing acquisition targets.
Q: Why is AI moving up the M&A agenda?
AI is changing how businesses generate value across software, engineering and professional services.
It is reshaping customer expectations, service delivery and the economics of expertise. That affects pricing power, operating costs, customer retention and future cash flows.
Our research shows its influence on deals. KPMG surveyed 149 senior M&A professionals at listed companies earning over US$100 million annually. Some 78% said AI was changing their acquisition assessment criteria.
Yet no accepted framework measures AI’s impact on enterprise value. Traditional due diligence can miss disruption that overturns an investment thesis.
Q: What exactly is AI defensibility and why does it matter?
AI defensibility measures how effectively a business can protect its competitive advantage and financial returns against AI disruption.
Resilience differs from adoption. A company might deploy AI extensively while remaining vulnerable to competitors using the same technology.
Businesses adopting little AI might still be protected by regulatory protection, proprietary expertise or deeply embedded customer relationships.
The question is not simply whether an acquisition target uses AI. It is whether AI strengthens or undermines the characteristics that make that business successful.
Q: Is there agreement about what makes businesses AI-defensible?
Not yet, creating uncertainty for investors.
Our research found that only two characteristics attracted majority support among dealmakers. Regulatory, compliance and security barriers were identified by 55%, while workflow integration received 52%.
Competitors cannot easily replicate established regulatory permissions or deeply embedded operational processes.
Other potential advantages include proprietary data, specialist expertise, switching costs, network effects and AI architecture. Their importance remains disputed.
That creates valuation risks. Two businesses operating in comparable markets could receive dramatically different assessments because buyers apply inconsistent assumptions about competitive resilience.
Q: What should dealmakers examine when assessing acquisition targets?
I would focus on questions affecting enterprise value.
First, could AI compress revenue by making existing services cheaper or easier to replicate? Buyers must understand whether productivity improvements translate into sustainable margins or simply encourage customers to demand lower prices.
Second, are efficiency gains being offset by rising expenditure on computing infrastructure, software and external AI providers?
Businesses sometimes use costly models for simple tasks. That undermines expected savings.
Third, could customers or competitors bypass the business entirely?
Consider advisory firms whose commercial proposition depends heavily on analysing and synthesising information. AI could commoditise parts of that offering, making proprietary knowledge, trusted relationships and regulatory credentials increasingly important.
Fourth, could AI fundamentally change how customers obtain value?
When electricity transformed manufacturing, some factories initially adapted existing processes rather than redesigning production around the technology.
Businesses face a similar challenge today. Simply adding AI to established workflows may deliver incremental improvements without capturing its transformational potential.
Finally, can the organisation innovate faster than competitors? That requires investment, appropriate talent, reliable data and the organisational ability to implement meaningful change.
Q: How should buyers determine whether a target is genuinely AI-ready?
Look beyond technology announcements and examine the operational foundations.
Reliable, accessible data is essential. Without it, even sophisticated AI systems struggle to deliver dependable commercial outcomes.
Buyers should establish whether AI is embedded in everyday operations or confined to experimental projects with limited business impact.
Talent, governance and management accountability matter equally. Organisations need people capable of deploying AI responsibly, measuring results and responding when systems underperform.
One revealing indicator is whether management understands where AI should not be deployed.
Not every activity requires an advanced model, and some processes are better handled through conventional automation.
Businesses demonstrating disciplined technology choices, measurable productivity improvements and consistently accurate outputs are generally better positioned than organisations showcasing numerous disconnected AI initiatives.
Q: What should dealmakers change immediately?
AI assessment needs to begin during acquisition screening rather than appearing as an additional exercise near the end of due diligence.
Buyers should evaluate both opportunities and threats. AI could accelerate growth, improve productivity and strengthen customer relationships. Equally, it could erode revenues, weaken margins or undermine established market positions.
Consistency is essential. Applying a common assessment framework across acquisition targets enables investors to compare businesses more effectively and identify differences in their exposure to disruption.
Sellers must also demonstrate why their competitive advantages will endure.
That means explaining how proprietary data, customer relationships, regulatory standing and workflow integration create barriers that competitors cannot easily overcome using readily available AI technologies.
Defensibility must be assessed dynamically. Advantages that appear durable today may weaken as models improve, computing becomes cheaper and competitors develop alternative business models.
A further complication is timing. A target may enjoy strong margins today while its customers are already experimenting with AI alternatives. Conventional financial statements will not necessarily reveal that exposure.
Dealmakers should therefore examine customer behaviour, contract renewals, product usage and the pace at which competing solutions are improving. They should also test scenarios rather than rely on a single forecast.
What happens if prices fall sharply, implementation costs rise or a customer brings previously outsourced work in-house? These questions can expose vulnerabilities that headline growth figures conceal.
Dealmakers therefore need to examine not just current capabilities but management’s capacity to adapt as technological and commercial conditions evolve, even before the transaction is completed.
Ultimately, AI is making competitive advantage more difficult to measure precisely when accurate valuation matters most.
The next decisive advantage in M&A may not belong to organisations deploying the most AI. It may belong to investors who understand its economic consequences better than their competitors.
