The question of whether AI will kill company moats is a provocative one, but it's not as simple as a yes or no answer. AI is forcing us to reconsider the nature of sustainable competitive advantages, but it's not going to destroy them all. Some moats will remain largely unaffected, while others will narrow or disappear. The key is to understand the composition of these moats and the potential risks and weaknesses they may face. AI is making many previously scarce skills, large teams, and years of expertise cheaper, faster, and more widely available. This could erode the capabilities that once helped distinguish software companies, creative businesses, and consulting firms. For example, writing software, producing advertising, and analyzing documents were once difficult and expensive, but AI is making these tasks more accessible. This doesn't mean these companies will become irrelevant, but it does mean that some of their unique advantages may be easier for competitors to replicate. Switching costs, which are often a barrier to entry for customers, are another factor to consider. AI can help with tasks like writing code, translating data, and building integrations, making it easier for customers to switch providers. This could potentially narrow the moat of companies that rely heavily on these switching costs. Brands are also at risk. A trusted brand helps customers make decisions, but what happens when an AI assistant makes those decisions for us? If an AI agent compares insurance policies, mobile phone plans, or retailers and chooses the best option, the primary customer relationship might belong to the AI platform, not the business. This could put pressure on brands and potentially erode their competitive advantage. However, some competitive advantages are more resilient than others. AI can create property websites, but it can't replicate the unique listings and audience of REA Group Ltd. AI can assist in medical research, but it can't quickly replicate CSL Ltd's plasma collection network, manufacturing capability, and accumulated know-how. Similarly, while AI can improve banking technology, it can't simply hand a new entrant like Commonwealth Bank of Australia its license, deposit base, and public trust. AI also can't manufacture scarce mineral deposits, prime locations, physical distribution networks, or genuine economies of scale. These more resilient moats are not invulnerable, though. A company might retain its network but lose control of the customer interface, or a trusted incumbent might keep its customers while finding its products easier to compare and its margins harder to defend. So, what should investors do? Instead of predicting the future, they should prepare by understanding the composition of the moat. This includes asking whether the company owns something genuinely scarce or merely possesses a capability that AI could commoditize. It's also important to consider whether switching costs are structural or just difficult and annoying. Investors should also assess whether the company owns the customer relationship or if an AI assistant could insert itself between the business and its customers. And, as always, valuation matters. A wonderful company can be a lousy investment if its share price assumes the moat will last forever. If the future of the moat has become less certain, investors should demand a larger margin of safety. In conclusion, AI is not the end of moats, but it might be the end of taking their permanence for granted. Investors should prepare for potential risks and weaknesses, so they can notice if and when they start to impact a company's profits or prospects. This proactive approach will help them make more informed investment decisions in an increasingly AI-driven world.
