Greg Cross: Customer Data Is the Most Defensible Moat in the AI Era

Is your customer data decaying? Greg Cross says it is. In his essay Customer Data: The Most Defensible Moat in the AI Era, published on his Substack in October 2025, he looks for inflection points with three questions: what's new, what's the same, and what's the same but different. The third, he says, is where the critical decisions get made.

This is our summary of his argument. The views and forecasts below are his.

What's new: AI has levelled execution

ChatGPT reached 100 million users in two months, and tools that once needed whole teams now need a prompt. In Greg Cross's view, that has democratised execution: Porter's traditional moats (economies of scale, proprietary technology, distribution advantages) are filling in.

But he argues AI has also created a new advantage: proprietary customer data that no competitor can replicate. His point is not simply that data is valuable. It is that trust decides how defensible the data is. The deeper the trust, the richer the data, and the richer the data, the stronger the advantage.

He sees three forces converging:

1. AI has made personalised experiences at scale economically viable.
2. Customers have begun looking for belonging and authentic connection.
3. Technology foundations now exist to make trust measurable at scale.

What's the same: trust still decides

Some things have not moved. Moats competitors can't cross still win. The 80/20 principle still holds: twenty percent of customers generate eighty percent of profits. Relationships create value that doesn't show on a balance sheet. People still decide on trust, not features. And methods for valuing relationships already exist.

What's the same but different: data that decays

Here he turns to the data businesses already hold. A customer relationship management system records purchases, clicks and demographics, and he argues these tell you less each quarter. Competitors run similar models on similar data, so AI removes whatever edge was left. Surface signals can't show why someone chose you.

Meanwhile, he says, relationship data lives in customers' lives, not in your systems. Your top 20% get treated like everyone else. Community work only earns budget once trust becomes measurable. In his words, businesses are measuring the outcome while missing the cause.

Transactional data decays, relationship data compounds

This is the distinction his essay rests on. Transactional data has a half-life, and he argues that AI makes it shorter every month. Relationship data compounds: trust deepens with every interaction. You can't buy it from data brokers or scrape it from public sources, and a competitor can't copy it.

Why he says to act early

For Greg Cross, timing is mathematics. A competitor that starts twelve months later starts behind the compounded value of those twelve months, and he expects the gap to grow exponentially rather than linearly. He warns that AI vendors want your data, not your success, that missing one budget cycle can leave a business twelve to eighteen months behind, and that the privacy frameworks governing this are being written now.

Two paths

He describes two paths. With trust infrastructure, a business gets predictive personalisation, network effects and relationship assets it can measure. Without it, its data decays and its best customers leave.

His verdict: the company that knows its customers in relationship terms, not just transactional ones, wins the AI era, because data is the new moat and trust is what makes that moat defensible. He closes by saying his own company has built infrastructure for this and is offering pilots.

The takeaway we draw from it: measure the cause, not just the outcome.

Read Greg Cross's full essay on Substack.