ai-powered marketing, ivan jerkic

The human factor in an AI-powered marketing

Reading Time: 3 minutes

Table of Contents

Ivan Jerkić

Ivan Jerkić

What do I do? I build ventures – from the concept to product-market fit. And then I scale them.

Forget keywords, demographics, affinities, and narrowed-down customer personas.

Google search ads work better if you ditch exact keywords and keep them broad, letting the machine learning algorithm detect users’ search intent. TikTok’s algorithm can distinguish strong purchase intent from affinity based on how the user behaves in the app. Meta’s machine-learning system predicts user behavior using only modeled data generated by opted-out users on iOS.

My point is that AI has advanced enough to overperform the human digital marketer in 10 out of 10 cases. After all, it has access to more data points in real-time than you will ever have. AI is better at predicting patterns, and our digital footprint is nothing more than a perfectly laid out behavior pattern. 

Even though you might think that it means we are playing on a leveled field and that anyone can set up a successful ad campaign powered by machine learning, growth marketing is more complex and individualized than ever.

We are not digital marketing or PPC specialists. We are teachers, doctors, pilots, and strategists.

We are working alongside AI systems to fill in the machine’s limitations: context, nuance, and creativity — leading to strategy.


Teachers train the system, both initially and later as needed to get optimum results.

Ad platforms’ machine learning systems collect much data but don’t understand your unique business case by default. AI needs additional input to know how your customers behave.

ai-powered marketing, ivan jerkic
Source: Unsplash

Rather than manually entering targeting parameters to create personas, you should provide high-quality business data and let the AI do its own thing. This data includes transaction history, user information, engagement data, and even some less-known sources that can fine-tune the algorithm, like weather data if your business is seasonal or sentiment data if your business depends on public opinion. We are calling this automation layering.

To make it neat, keep all the data on the customer data platform (such as Segment or Lytics), where you can efficiently work with it and decide what to push to a particular ad network.


Doctors diagnose problems and define actions to resolve them.

In the beginning, when we don’t have enough data, AI has issues with context and nuance. This is the phase when humans should intervene and often look after search terms and ad placements.

These problems are hard to pinpoint with another automation, but you can set up custom parameters to detect anomalies in performance.

For example, sometimes ads appear on placements that are either out of context for the message we try to convey or on placements using dark UX patterns that deceive users into clicking on ads. We can detect both issues by the fair share of clicks and a reasonable click-through rate but with bad on-site engagement metrics.


Pilots monitor performance and take control from the AI when necessary to put it back on the right track.

Seasonality and critical offline events can throw performance off-balance. Just as pilots would do in a similar situation, advertising specialists need to turn off the “autopilot” and steer the system in the right direction.

ai-powered marketing, ivan jerkic
Source: Unsplash

Sometimes you need to adjust the target return on ad spend or change a bid strategy. On the other hand, sometimes it would be necessary to re-adjust communication due to a new competitor entering the market.

You can set up automated rules that trigger alerts when something goes wrong, i.e., outside your set parameters to watch after. This will help you recognize the moment when it’s necessary to take action.


Strategists define the goals and means to provide context, nuance, and creativity to the whole project. Also, they draw meaningful business insights from the performance and results.

A strategist is an essential role your growth marketing team needs to have to have a shot at success. The sooner you embrace this new paradigm and fully utilize AI’s operational capabilities while focusing on what machines cannot accomplish (as of yet) and what matters the most — strategy and creativity — you will see some great results!


Unlevel the Playing Field: The Biggest Mindshift in PPC History by Frederick Vallaeys

Share this post

Latest posts

scaling, ivan jerkic
Growth Marketing

The scaling trap

Reading Time: 2 minutes In general, many startups and small businesses tend to over-plan their scaling. Scaling is excellent; it will eventually bring money to your business – but it’s pointless to plan your growth routes before you even have a functional car (if you get the analogy).

Read more
can AI be creative?, ivan jerkic

Can AI be creative?

Reading Time: 3 minutes So here it comes – the overarching question many people in tech debated over the last few months. Can AI be creative? In this post, I will explain my stance and see how some of the most prominent AI systems respond when prompted about their creativity.

Read more