EVOLVE connects NED’s, board-level leaders, senior advisors and fractional executives with growth-stage commerce companies. Our events are one expression of that work.

We host expert sessions, in-person meetups, and small, closed dinners across key global markets. These gatherings are designed to support ongoing conversations, not one-off appearances.

Now, onto this week’s newsletter.

This week marked our 100th EVOLVE expert session.

The guest was Mihai Barsan, founder of The Engine, former brand leader within Ursus, SABMiller and Procter & Gamble. His career has unfolded inside high-volume consumer environments where distribution is constrained, budgets are scrutinised, and creative output must convert into measurable commercial movement.

He operates at the intersection of storytelling, behavioural economics and brand systems.

The debate we hosted was not about tools. It was about discernment.

We opened with a simple provocation:

If AI can generate infinite variations of content, does originality still matter?

Mihai’s answer was immediate:

This distinction framed the entire discussion.

AI lowers the cost of production.
Distribution platforms lower the barrier to publishing.

But abundance does not eliminate the need for judgment.

In fact, it intensifies it.

The instinctive explanation for the current anxiety is saturation.

Feeds are crowded. Formats multiply. Attention fragments. Every scroll surfaces another branded narrative engineered for velocity. The conclusion appears straightforward: breaking through is harder because there is simply more to compete against.

That framing misses the structural shift.

Advertising has survived abundance before. Desktop publishing lowered barriers to print. Social platforms dissolved distribution gatekeepers. Programmatic buying enabled infinite banner variations. Each cycle expanded output dramatically. Each cycle triggered the same complaint about noise.

Brands still built memorability.

The difference today is not quantity. It is friction.

In previous eras, even trivial campaigns required coordination. Production imposed constraint. Budgets were finite. Timelines were fixed. The effort involved in making something forced teams to debate before publishing.

Generative systems remove that constraint.

A marketer can describe a concept and receive a finished storyboard within minutes. Iteration becomes conversational. Adjustments that once required negotiation now require a prompt.

The result is not only more content. It is faster replication.

This alters internal behaviour. When production is expensive, scrutiny happens before launch. When production is cheap, scrutiny often shifts to performance metrics after launch.

The discipline moves downstream.

The Romanian commercial Mihai criticised did not fail because it was lost in a crowded feed. It failed because it lacked orientation. The narrative beats were visible. The connective tissue was weak. The execution appeared complete. The message did not anchor.

The language model that evaluated it recognised the presence of a hero and a resolution. It did not assess whether the audience would care.

That is the distinction.

Saturation amplifies everything. It does not create emptiness. It reveals it.

If audiences are inundated with interchangeable narratives, their tolerance narrows. Repetition becomes easier to ignore. What persists under those conditions is not volume. It is clarity.

AI introduces fluency at scale. It assembles language and imagery convincingly. It references archetypes effortlessly. Fluency, however, is not distinctiveness.

The risk for organisations is confusing the two.

When every idea can be executed immediately, fewer ideas are interrogated deeply. The natural pause that once existed between concept and production disappears. Without deliberate standards, teams default to motion.

Platforms reward activity. Algorithms reward frequency. Neither reward coherence directly.

The responsibility to maintain coherence remains internal.

This is why saturation is not the central threat. Indiscriminate acceleration is.

The flood is manageable. The erosion of discernment is not.

The experiment Mihai described was simple. He ran two advertisements through a language model. One was emotionally coherent. The other was the Romanian AI-generated spot he had criticised. The model rated both highly on structural storytelling markers.

It recognised narrative components. It detected a protagonist, tension, resolution. It identified completeness.

It did not assess meaning.

Large language models are trained on pattern recognition at scale. They measure fluency and structural consistency. They do not experience consequence. In marketing environments, that gap matters.

McKinsey’s 2023 report The Economic Potential of Generative AI identifies marketing and sales as among the functions most immediately affected by generative systems, particularly in content creation and personalisation. The productivity upside is clear. More assets. Faster cycles. Lower marginal cost.

Productivity, however, does not equal distinction.

WARC’s 2023 analysis of global advertising effectiveness shows a sustained decline in long-term brand-building impact in markets dominated by short-term performance optimisation. When execution becomes easier and testing cycles shorten, immediate metrics tend to crowd out slower equity accumulation.

Generative systems intensify that bias.

If a team can produce ten variations of a message in a day and optimise against click-through or conversion, the structural incentive shifts toward surface responsiveness. Narrative coherence becomes secondary to measurable lift.

The risk is subtle. Structure becomes abundant. Meaning does not.

When Mihai criticised the Romanian spot, his concern was not technical polish. It was detachment. The narrative scaffold was present, but it was not anchored in product truth. It borrowed emotional rhythm without earning it.

AI reproduces archetypes effectively. It struggles to embed them in lived brand reality.

The distinction is economic. Brands that generate impressions without attachment weaken their pricing power over time. The erosion is gradual. Dashboards rarely capture it immediately.

When structure becomes commoditised, discernment becomes the scarce resource.

That is the pressure point

During the session, Mihai introduced a ratio that reframed the discussion. Roughly 5 to 25 percent of advertising demand, he argued, is anchored in the present. The remaining 75 to 95 percent sits in the future

His point was not numerical precision. It was orientation.

Most marketing activity is evaluated against immediate signals. Clicks. Conversions. Engagement rates. Short-term revenue attribution. These metrics are tangible and reportable. They fit quarterly rhythms.

Brand demand accumulates differently. It builds in memory structures, emotional associations and category linkage. It does not spike in dashboards the week after launch. It compounds quietly.

The imbalance between short-term optimisation and long-term demand creation is not new. The IPA’s long-running effectiveness studies, particularly the work of Les Binet and Peter Field, have shown that campaigns weighted too heavily toward activation tend to underperform over longer horizons. Brand investment produces disproportionate returns over time, but the lag makes it vulnerable to budget reallocation.

Generative systems accelerate the temptation to over-index on immediacy.

If content can be produced quickly and tested rapidly, the feedback loop tightens around short-term signals. Variations that perform incrementally better in the moment are scaled. Those that do not are discarded. The system favours responsiveness.

That does not inherently undermine brand-building. It does, however, bias organisational attention.

Mihai’s concern was that the structural shift in production cost could distort evaluation frameworks. If teams begin to treat advertising primarily as an optimisation exercise, they risk neglecting the slower layer where brand equity forms.

The irony is that generative AI could enable stronger long-term storytelling. Lower production cost frees budget for experimentation. Iteration could deepen narrative exploration.

Whether that potential is realised depends on governance.

If performance metrics dominate decision-making, content volume will rise while distinctiveness erodes. If long-term demand remains a defined objective, generative systems become tools rather than drivers.

The economic question is not whether AI reduces cost. It does. The question is whether organisations preserve the discipline to invest in what cannot be measured immediately.

Demand does not disappear when ignored. It shifts toward competitors who remain recognisable.

That is the strategic tension

Content does not exist in isolation. It circulates inside systems designed to reward activity.

Most major platforms optimise for engagement velocity. Time spent. Interaction rate. Recency. The logic is straightforward: more activity increases monetisable attention. The system rewards frequency.

Generative AI aligns naturally with that incentive structure. If content can be produced at negligible marginal cost, output increases. If output increases, the probability of triggering engagement rises statistically.

This creates artificial volume.

Artificial volume does not necessarily mean artificial impact. It means the production constraint is no longer economic. It is editorial.

That shift alters competitive dynamics.

In previous cycles, distribution bottlenecks filtered content implicitly. Television inventory was finite. Print pages were limited. Even early digital placements required media negotiation. Scarcity moderated supply.

Digital platforms eliminated distribution scarcity. Generative systems eliminate production scarcity.

When both constraints weaken simultaneously, the burden of differentiation moves entirely to positioning and narrative coherence.

There is also a second-order effect.

If generative systems are trained on the corpus of platform content, and platforms reward engagement patterns that generative systems can replicate, a feedback loop forms. Content becomes optimised for algorithmic recognition. Distinctiveness becomes statistically rarer.

In financial markets, regulators have warned about model convergence risk, where similar algorithms amplify volatility. In creative markets, convergence produces sameness.

Mihai’s discomfort with the Romanian advertisement was not aesthetic elitism. It was recognition of convergence. The ad resembled other ads. It satisfied formal criteria. It did not assert identity.

Platforms cannot correct for that. Their mandate is engagement, not brand coherence.

Which returns responsibility to operators.

Volume is not inherently harmful. It becomes harmful when it substitutes for clarity.

The temptation, especially under pressure, is to respond to crowded feeds with more output. More variants. More experiments. More distribution.

The alternative is slower and more difficult. Define the boundary conditions of the brand. Use generative systems to refine expression within those boundaries. Resist the drift toward homogeneity.

Artificial volume is easy. Coherent volume requires governance.

That is where advantage will sit.

When production costs fall, standards become cultural rather than economic.

In earlier cycles, constraint imposed discipline. Budget approvals functioned as filters. Now the filter must be intentional.

This is an executive issue.

Creative teams can generate options at scale. Performance teams can test rapidly. Agencies can automate execution. None of those capabilities determine whether the output strengthens the brand.

That decision sits higher.

Governance does not mean restricting experimentation. It means defining non-negotiables. What must remain consistent. What cannot be diluted. What emotional territory belongs to the brand.

Without those guardrails, generative systems default to statistical familiarity. Familiarity performs adequately in the short term. It rarely builds attachment.

Mihai’s critique was rooted in this distinction. The Romanian commercial followed structure. It did not demonstrate authorship. Authorship requires a point of view.

In organisations where speed becomes the dominant virtue, authorship is often the first casualty.

The strategic risk is quiet erosion. Nothing collapses. Metrics remain acceptable. Over time, differentiation narrows.

Governance is the counterweight.

Lower cost production increases optionality. It also increases responsibility.

The brands that navigate this transition well will not be those producing the most content. They will be those who preserve narrative integrity while using new tools with precision.

That requires executive discipline.“Trust is bigger than outcome.”

In enterprise environments, that is not hyperbole. When liability exposure is high, trust determines who is invited into the room long before output is measured.

The session did not conclude with a verdict on AI.

It returned to a more practical question: what do operators control?

They do not control the volume of content circulating in the market. They do not control platform incentives. They do not control the pace of generative improvement.

They control standards.

Mihai’s critique of the Romanian commercial was not technological. It was editorial. The ad looked complete. It was not coherent. That difference is subtle in the moment. It compounds over time.

Generative systems will continue to reduce the cost of execution. They will become better at mimicking tone. They will refine pacing, visual continuity, linguistic fluency.

None of that guarantees that a brand means something.

Meaning emerges when narrative, product truth and audience expectation align. That alignment cannot be automated fully because it depends on judgment. Judgment sits with people accountable for long-term value.

The flood of content will intensify. Some of it will be effective. Much of it will be forgettable.

The dividing line will not be technical sophistication. It will be discipline.

This hundredth session did not celebrate creative nostalgia. It examined exposure. When structure is easy to generate, substance becomes the scarce asset.

The organisations that recognise this early will treat generative tools as amplifiers, not authors.

That distinction will not trend loudly.

It will show up in pricing power, recall and resilience.

Quietly.Beyond Agencies

This conversation marked the hundredth structured EVOLVE expert session. EVOLVE convenes board-level leaders and senior operators in ongoing, closed discussions focused on structural change inside commerce and adjacent systems. The purpose is not trend commentary. It is to examine shifts while they are forming, before they harden into consensus.

Curated. Personable. Global.
That is EVOLVE

If you’d like to learn more how to partner with, please email us at
[email protected] or [email protected]

Podcasts & Media We Support