Marketing in the AI Era: Yuliya Lennox on Smaller Teams, Bigger Budgets, and Human Judgment

AI has made it possible for one marketer to do work that previously required an entire team. Creative production is cheaper, websites can be built through conversations with coding agents, and managing large advertising budgets no longer necessarily requires dozens of specialists. But when everyone has access to the same tools, what actually separates a good marketer from someone who knows how to write prompts?

In the latest episode of the Retention Podcast, Oleg Lesov discusses this question with Yuliya Lennox, CMO in an AI project at Genesis. Her career includes Coca-Cola, BetterMe, Replika AI, Higgsfield, and Solidstarts, giving her experience with established global brands, rapidly growing consumer apps, and AI products searching for their place in the market.

Their conversation covers the changing economics of marketing teams, the relationship between performance marketing and brand building, the challenges of finding product-market fit, and the skills marketers need as AI takes over more operational work.

One thing becomes clear throughout the discussion: AI is changing how marketing gets done, but experience, creativity, and independent thinking remain central to making it work.

Why Marketing Needs a System

Having worked with several successful companies, Yuliya has noticed one common characteristic: strong leadership. The companies may operate in completely different industries, but they tend to have leaders who create momentum, communicate a clear direction, and make people believe in what they are building.

However, she makes an interesting distinction between leadership and marketing execution.

While leadership can depend heavily on individual personalities, marketing should not. A successful marketing operation needs a structure that continues producing results without depending on one exceptional person.

For Yuliya, that means establishing a monetization strategy, understanding which creative approaches work, building repeatable processes, and creating a system that can scale.

This becomes particularly important as AI changes the relationship between team size and output.

A few years ago, producing hundreds of advertising creatives required motion designers, graphic designers, copywriters, and marketers working together. Today, AI tools can generate large volumes of creative materials at a fraction of the previous cost.

Yuliya describes a scenario where one person can spend approximately $200 on tools and produce around 1,000 creatives. That doesn't mean all those creatives will perform, but it illustrates how dramatically production economics have changed.

The implications for team structure are significant.

Yuliya believes small teams can sometimes outperform larger organizations because they avoid the coordination problems that come with specialization. When one marketer understands the product, creative performance, audience, and acquisition strategy, decisions can happen much faster.

In a larger team, that same decision might require several meetings and coordination between departments.

Of course, managing hundreds of millions in advertising spend may still require a larger organization. But team size should reflect actual operational needs rather than assumptions about how many people a marketing department is supposed to have.

Can One Person Manage $100,000 in Daily Ad Spend?

One of the more provocative ideas in the conversation is Yuliya's belief that a single marketer could manage approximately $100,000 in daily advertising spend.

That translates into roughly $3 million per month.

Importantly, she presents this as her assessment of what is possible with today's tools and market conditions, rather than a scale she has personally managed alone.

Her argument is that the biggest challenge isn't necessarily operating the advertising account. It's finding a product and creative strategy that can support that level of spending profitably.

Once a company has strong product-market fit and successful creatives, scaling becomes a different kind of problem.

Yuliya recalls that Replika operated with a very small user acquisition function even while managing substantial advertising activity. The example reinforces her belief that successful performance marketing doesn't always require a large department.

However, she also warns against becoming too fascinated with hypothetical scaling numbers.

For a startup spending $100 per day, worrying about how to manage $100,000 per day is premature. The immediate priority should be finding an audience, testing positioning, and proving that acquisition can generate sustainable revenue.

The distinction between finding product-market fit and scaling an established product is important because the two stages require different approaches.

When the product is already successful, marketers can concentrate on improving and expanding a working system. When the company is still searching for demand, almost every assumption remains open to challenge.

And that uncertainty can be much harder to manage than a large advertising budget.

Finding Product-Market Fit Is Harder Than Scaling a Proven Product

Yuliya's experiences at Replika and Higgsfield illustrate how different these stages can feel.

When she joined Replika, the product had already demonstrated substantial demand. It had experienced viral growth, attracted organic users, and established a recognizable proposition.

Performance marketing could build on that foundation.

Higgsfield was a different experience.

During her time there, the company was still searching for the right market position. The team tested different directions, reconsidered product decisions, and had to abandon ideas that initially seemed promising.

At one stage, the company believed a mobile app was the right format. Eventually, the team reconsidered whether an app was actually the best way to deliver its technology.

Yuliya describes that period as difficult because even experienced marketers cannot guarantee that a promising idea will succeed.

The same uncertainty appears in smaller product experiments.

She recalls working on an English-learning app where the team believed allowing users to upload their own books would be an attractive feature. It sounded useful, especially for people who wanted to learn through familiar literature.

But users showed little interest.

The feature required effort from customers and did not deliver the expected results.

Experiences like these have shaped Yuliya's approach to marketing. Professional experience can improve the quality of a hypothesis, but it cannot replace testing.

The most successful ideas are not always the most sophisticated ones. Sometimes a relatively simple feature or creative concept performs better than something the team spent weeks developing.

For early-stage companies, the ability to recognize that difference quickly is essential.

Performance Marketing and Brand Building Should Work Together

Yuliya's career began at Coca-Cola, where brand marketing played a major role. Later, working in performance-driven technology companies, she became much more focused on measurable acquisition results.

That transition initially made her skeptical of brand marketing.

When you're accustomed to measuring clicks, conversions, and revenue, it can be difficult to justify activities that don't produce an immediate financial return.

Over time, however, she changed her perspective.

One example comes from watching Duolingo's development. Yuliya remembers a period when another language-learning product she worked with was competing closely with Duolingo. The company later saw Duolingo strengthen its marketing, expand its reach, and develop a much more recognizable brand.

For Yuliya, this demonstrated that brand recognition can become an important competitive advantage, particularly as a company grows.

Higgsfield provided another example. When the company first started attracting attention, many people outside the AI industry were unfamiliar with it. Later, viral content helped introduce the brand to a much broader audience.

That recognition has value beyond immediate conversions. It influences how customers, partners, potential employees, and investors perceive the business.

Still, Yuliya doesn't recommend prioritizing brand building at the expense of performance marketing.

For a startup without substantial funding, generating revenue and finding product-market fit remain essential. Brand activities should support that process rather than consume resources the business needs to survive.

Her approach is to start testing acquisition early while establishing basic brand channels in parallel.

Even something as simple as maintaining an active Instagram account can help capture attention generated by paid advertising. People who discover a product through ads may visit its social profiles before deciding whether to trust it.

Without those channels, companies can miss opportunities to turn paid exposure into longer-term recognition.

The challenge becomes finding the right balance between immediate performance and the brand a company wants to build over several years.

Why Brand Guidelines Can Limit Creative Testing

Brand consistency becomes particularly complicated when performance marketers want to experiment.

Yuliya shares an example from her time at Solidstarts, a product connected to children's nutrition.

She noticed a creative trend featuring AI-generated characters resembling children made from fruit. The format was visually engaging, and she decided to test something similar.

The creative produced acceptable performance results. However, the reaction from colleagues and people online was negative. The problem wasn't necessarily the advertising metrics. It was the mismatch between the creative and the product's identity.

Parents using a nutrition-related product expect credibility and responsibility. A playful AI-generated character might attract attention, but it can also make the company appear less serious about a sensitive subject.

That experience highlighted a tension between creative freedom and brand trust.

Performance marketers often want to test unusual hooks, visuals, and messages because unexpected ideas can outperform conventional advertising. Brand teams, meanwhile, need to consider how those experiments affect customer perception.

Neither perspective is automatically wrong. A company without an established audience can often experiment more freely because it has fewer expectations to manage. An established brand has more to lose when a creative contradicts what customers believe about it.

This becomes especially relevant as AI makes it possible to generate hundreds of creative variations almost instantly.

The production cost may be low, but the potential reputational cost of publishing the wrong message hasn't disappeared.

AI Makes Creative Production Cheaper, Not Automatically Better

The ability to generate advertising content at scale is one of the biggest changes Yuliya has witnessed.

Tasks that previously required significant production budgets can now be completed using relatively inexpensive AI tools. Video generation, image creation, copywriting, and website development have all become much more accessible.

But Yuliya is skeptical of the idea that creative teams can already be replaced completely.

She describes seeing a presentation about an automated Instagram account where AI agents handled the entire content process. The creator was proud that no human team was needed.

The problem was the output.

In Yuliya's assessment, the content looked obviously AI-generated and lacked the quality needed to perform effectively.

This illustrates an important difference between producing something and producing something worth publishing.

AI can generate hundreds of assets, but volume alone doesn't guarantee that those assets will resonate with an audience.

Marketers still need to understand customer motivations, recognize emerging trends, evaluate creative quality, and connect advertising concepts to a product's positioning.

For now, Yuliya prefers working with designers who understand how to use AI rather than relying entirely on automated creative systems.

She expects the technology to continue improving and believes more production tasks will eventually become automated. However, the value of experienced creative professionals increasingly depends on how they use those tools and the judgment they bring to the final result.

The question is no longer simply how many creatives a team can produce. It's how effectively the team can identify which ideas deserve to be tested and which outputs are good enough to represent the brand.

AI-Native Marketing Teams Need People Who Can Think Independently

The conversation eventually turns to a problem that many companies are encountering as AI becomes part of everyday work.

How do you evaluate a marketer when AI can complete much of the assignment for them?

Yuliya shares an example from her hiring experience.

She gave a candidate a relatively simple dataset and asked them to analyze it. The task could have been completed using basic spreadsheet functionality, including a pivot table.

Instead, the candidate submitted a lengthy analysis generated through ChatGPT.

When Yuliya asked the candidate to explain the numbers and demonstrate the analysis, they couldn't do it.

The issue wasn't that the candidate had used AI. It was that they couldn't explain or verify the result.

Yuliya believes this distinction will become increasingly important.

Marketers who simply accept the first recommendation from an AI model aren't necessarily adding much value. If a person cannot explain why a campaign should be scaled, why a creative should be tested, or what the numbers mean, the company is effectively relying on the model rather than the employee.

She wants marketers who have their own opinions.

In performance marketing, this includes developing an understanding of advertising platforms through experience. Marketers learn to recognize unusual performance patterns, understand when a campaign needs more time, and make decisions that may not be obvious from a single report.

AI can help analyze information and suggest possible actions, but experienced professionals still need to evaluate those suggestions.

Oleg shares a similar example from hiring at Reteno, where a candidate was asked to compare Reteno with competing platforms. Rather than exploring the actual products, the candidate relied entirely on AI-generated information.

Both examples point to the same conclusion: using AI is not a substitute for understanding the work.

For marketing teams, the strongest combination is likely to be professionals who understand their discipline and can use AI to move faster.

AI Is Changing Jobs, but It Is Also Changing Expectations

Yuliya has already seen AI influence hiring decisions and team structure.

When building a marketing team, she increasingly questions whether every traditional production role needs to be filled. A second graphic designer or another motion designer may be less necessary when existing team members can use AI to handle part of that workload.

However, she also expects specialists to adapt.

She recalls working with a motion designer who had been excellent several years earlier. When they collaborated again in 2026, she noticed that the person's approach had barely changed despite the availability of new AI tools.

For Yuliya, that was a problem.

The expectations of creative professionals have shifted. Experience remains valuable, but the ability to incorporate new technology into existing workflows is becoming part of professional competence.

Yuliya has built websites and marketing funnels using Claude, reducing the need for some development work. She sees these capabilities as particularly valuable for small teams that want to test ideas without building a large technical organization.

But the productivity gains have an unexpected consequence.

Instead of reducing workloads, AI can create pressure to accomplish even more.

Oleg describes experiencing this himself. With coding agents and new AI tools making so many tasks possible, it becomes tempting to spend every available hour experimenting, building, and learning.

Yuliya recognizes the same pattern. The technology may reduce the time required for individual tasks, but it also raises expectations about what one person should be able to accomplish.

This creates a new challenge for leaders: deciding which opportunities deserve attention when the technical limitations that previously prevented experimentation are disappearing.

Could a One-Person Company Become a Unicorn?

Yuliya believes that AI will eventually make it possible for a single person to build a billion-dollar company.

She sees the necessary tools becoming increasingly accessible.

A founder can develop a website with AI, generate advertising creatives, create social content, build marketing funnels, and potentially even develop physical products using technologies such as 3D printing.

Activities that previously required several departments can increasingly be handled by one person working with software and AI agents.

However, she doesn't believe that automation alone creates a valuable business.

A one-person company still needs an exceptional idea, a real customer problem, and a product that people are willing to pay for.

Simply building another AI wrapper or task-management tool is unlikely to create a billion-dollar business without meaningful differentiation.

This is an important qualification because AI has made experimentation easier for everyone.

Lower development and production costs create opportunities for small founders, but they also increase competition.

The ability to execute quickly becomes less distinctive when competitors can do the same thing.

Understanding customers, choosing the right market, and developing a product with genuine demand remain essential.

For Yuliya, AI expands what an individual can attempt. It doesn't eliminate the need for strong business judgment.

In a Nutshell

Yuliya's experience across Coca-Cola, BetterMe, Replika, Higgsfield, Solidstarts, and Genesis shows how much marketing has changed while some of its fundamentals have remained remarkably consistent.

AI has dramatically reduced the cost of creative production, accelerated website and funnel development, and made it possible for smaller teams to handle more work. A single marketer can potentially manage budgets that once required much larger departments, provided the product, creative strategy, and acquisition economics are already working.

But cheaper execution doesn't make finding product-market fit easier. Higgsfield's early experiments and Yuliya's experience with unsuccessful product features demonstrate that even experienced teams regularly misjudge what customers want.

Brand building also remains important, particularly for companies planning to compete over the long term. Performance marketing can validate demand and generate revenue, while a recognizable brand creates advantages that are harder to measure through individual campaigns.

Perhaps the most important change concerns the role of the marketer.

As AI takes over more execution, companies need people who can evaluate information, understand customer behavior, recognize good creative ideas, and make decisions independently.

The marketer who simply follows an AI recommendation may become less valuable. The marketer who knows when that recommendation is wrong becomes more valuable.

And while AI makes it possible to do more with fewer people, it also makes prioritization harder. When almost anything can be built or tested, knowing what deserves attention becomes a competitive advantage.

The future of marketing may involve much smaller teams and much more automation. But according to Yuliya, the people who succeed will still be the ones who understand the business, think for themselves, and know what they're trying to achieve.