MartechARTICLE

As generative AI becomes a commodity, competitive advantage moves back to those who think differently

As generative AI becomes a commodity, competitive advantage moves back to those who think differently
Image: Vanessa Guerra
Article · Martech
Reading
5 min
Tables
1
In short

A column by Reportei's CMO on creativity and AI exposes a bigger strategy problem: when the tool stops being an advantage, the competitive moat needs to come from somewhere else.

The question is about marketing. The answer that matters is about strategy

Renan Caixeiro, co-founder and CMO of Reportei (a Brazilian marketing-reporting platform), published a column in Exame (a Brazilian business magazine) on November 22, 2026, asking what still sets an idea apart when everyone uses generative AI to produce. His answer is repertoire, critical judgment, and authenticity. For whoever leads a creative team, that's a correct answer. For whoever signs off on the company's budget, the right question is different: if the tool stopped being an advantage, where does the competitive moat I was selling to investors two years ago now stand?

That's the conversation missing from boardrooms. A typical board approves licenses for generative AI tools, a code copilot, an image generator, a design assistant, half a dozen vertical copilots, expecting a productivity gain. It gets one. But productivity distributed equally among competitors isn't an advantage, it's a market floor. When every company in a sector speeds up by the same amount with the same tool, that speed stops showing up in the quarterly results as a differentiator and starts showing up only as a cost reduction, which anyone can copy within a quarter.

Where Caixeiro gets it right, and why this is product data, not just creative data

Caixeiro writes something worth quoting because it sums up the central point well:

Technology can reduce execution time, but it doesn't, on its own, create a reason for a brand to be remembered.

Renan Caixeiro, co-founder and CMO of Reportei

Translated into P&L terms: fast execution lowers production cost, but low production cost isn't what sustains margin when the competitor has the same cost. What sustains margin is the reason the customer chooses you instead of the cheaper substitute, and that reason never lived in delivery speed.

The same mechanism shows up in product. When a software feature becomes replicable by any competitor with access to the same model, what sustains a company's value stops being the isolated feature and becomes whatever can't be copied through feature parity: a plugin community, usage habits entrenched in a product team, network effects, switching costs.

When the feature becomes a commodity, the margin (or the valuation) that sustained the business should never have depended on it alone.

Where AI commoditization has already happened, and what it doesn't solve

It's worth separating what generative AI has already leveled from what it still doesn't touch, because confusing the two columns is the most expensive mistake I've seen a founder make in a product committee.

LayerWhat AI has already commoditizedWhat remains a human choice
ProductionGenerating a draft, image, script, or prototype in minutesDeciding which problem is worth solving
ReferenceInstant access to patterns and public dataChoosing references from outside your own industry
DeliveryVolume of content and variations per campaignConsistency between what the brand promises and what operations deliver
CostPrice per unit of creationJudgment on when perfect execution gets in the way of the message

Murilo Gun, an entrepreneur and speaker cited in Caixeiro's column, argues that good ideas depend on the quantity and variety of accumulated references, not on the tool used to express them. For a founder, this becomes a hiring and culture issue: a team that consults the same sources, follows the same creators, and uses the competitor's prompt will converge on the same result, no matter how good the model is.

The counter-argument I take seriously

There's a counterpoint that can't be ignored: is repertoire and authenticity really the moat, or is it comfortable rhetoric for whoever lacks proprietary data or distribution?

In practice, companies with a real competitive advantage almost always also have a hard-to-replicate asset beyond the team's repertoire: an exclusive database, deep integration with the customer's workflow, network effects, or switching costs. A collaboration platform with strong network effects has this in the form of a plugin community and entrenched product-team habits. A fintech that trains a credit-risk model on its own transactional history has this in the form of data competitors can't access.

Creativity and repertoire, on their own, don't guarantee defensibility against a competitor with capital and access to the same language model. Caixeiro's argument is necessary, but not sufficient: it explains why generic content loses relevance, not why a company keeps its margin when a competitor copies its creative strategy within a week. The complete answer combines both things, human repertoire to decide what to do and a proprietary asset to protect what was made.

The question I'd bring to the committee

Instead of asking how much the company is investing in generative AI, I'd ask the CTO and the CMO, in the same meeting, two specific things.

  • Which data, relationship, or usage habit does the company accumulate that a competitor can't replicate by buying the same AI license we bought?
  • At what point in the creative process does the team still form its own opinion before opening the tool, and at what point is it already just accepting the model's first answer?

Vinicius Gambeta, founder of Agência de Bolso, notes in the same column that the first answer tends to be the most obvious one, because it's closest to the most common reference available to the model. If the AI's first answer has become your company's final product, the competitor using the same model arrives at the same place, and the decision about where to invest the differentiation budget needs to move out of the marketing department and up to the boardroom table, because that's where it's decided which proprietary asset is worth building while the tool remains free for everyone.

Translated from the Brazilian Portuguese original · Read the original

More from Vanessa Guerra
View profile →