Johanna Frösén: Head-to-Head Interview

As marketing automation evolves into autonomous AI agents, how fundamentally will this change the way businesses plan, execute and manage marketing campaigns?

“Part of the more mundane marketing tasks will be automated, but the long-term strategic planning remains in human marketers’ hands. AI agents will not replace the need for a solid marketing strategy; instead, they expose the lack of one. Marketers’ role becomes increasingly about defining objectives, standards, guardrails, and acceptable trade-offs for AI-automated decisions, and less about implementation. Campaign planning and implementation become increasingly continuous as constant market testing and pivoting by AI agents become more commonplace.

“The underlying logic of marketing being fundamentally about exceeding competitors in understanding and fulfilling customer needs remains unchanged, whereas the tools of how to achieve this in practice are evolving perhaps faster than ever. Analysing diverse market data for enhanced customer insights, as well as translating this analysis into tailored offerings and communications become increasingly affordable and hence accessible to all. The bar for what constitutes effective planning, execution, and management or marketing campaigns is raised overall as autonomous AI agents become more prevalent.”

Which marketing decisions can businesses safely delegate to AI agents, and where should human oversight and approval remain essential?

“AI agents work best for decisions that follow predictable trajectories, come with observable short-term impacts, are easily reversible, and carry relatively low risks.

“AI works best in contexts where past performance is likely a good predictor of the future; less so in contexts where, for instance, completely new technology is introduced to the markets. Already before the recent surge of AI, the long-increasing data-drivenness in marketing has focused marketers’ attention on short-term impacts: reach, conversions, and sales. While AI-enhanced marketing analytics can improve many decisions by pairing creativity with hard market facts, data-drivenness is not without risks: an overfocus on short-term value capture may come at the cost of long-term brand building and customer performance. Hence, today’s big challenge is balancing analytics- and AI-informed insights with human long-term vision and brand strategy.

“Human oversight should also increase with the risk attached to the decision at hand: high involvement decisions that are irreversible, carry large implications in terms of the brand or the firm’s financials, or are ethically sensitive are best kept in human hands.

Can autonomous AI genuinely improve campaign performance, customer acquisition and marketing ROI, or are businesses overestimating its commercial potential?

“AI is particularly good at combining pieces of information from different sources, balancing attention between offerings, channels, and communications, as well as identifying patterns in customer behaviour. Gains are real in performance marketing, where feedback is quick and measurable: optimising campaign elements, identifying the most profitable target groups, or getting the highest bang for the buck invested in individual campaigns or communications.

“AI agents optimise what they can see, which tends to be clicks and conversions they can attribute to themselves. Marketing, however, is not only about short-term optimisation: differentiation strategies, long-term competitive advantage, and brand and customer relationship management all require balancing short-term optimisation with long-term business goals. Also, when competitors have access to the same agents, efficiency gains often get competed away. Thus, strategic decisions with long-term market implications are still best kept in human hands.”

What happens when autonomous marketing agents make poor targeting decisions, publish inappropriate content or overspend advertising budgets, and who should ultimately be held accountable?

“Accountability always rests with the human decision-maker. AI is subject to a variety of biases and hallucinations that may misguide decision-making. Hence, AI-based insights should be used more as input than as the outcome of any decision-making, and automated processes should be carefully limited to tasks that can be clearly structured, bounded, and therefore confidently trusted with AI.

“In the eyes of the customer, the marketer is always responsible for the customer experience: even if a service failure would be factually caused by the behaviour of a supplier, subcontractor, or AI agent, it is the brand that carries the responsibility for the customer relationship, as well as the related reputational risk.”

How can businesses maintain brand consistency, customer trust and regulatory compliance when AI agents independently create, personalise and distribute marketing content?

“Brand strategy, differentiation, and targeted competitive advantage form the basis of customer trust and consistency and are all decisions that require human judgement. Tailoring content to specific audiences, creating texts and visuals to communicate selected messages, and identifying appropriate channels to reach them are less risky decisions that can be better outsourced to AI. Consistency used to come from people who knew the brand; now it must be written down precisely enough for agents to follow.

“In all AI-based decisions, compliance with local regulation, such as the General Data Protection Regulation (GDPR; including limits on automated decision-making and profiling), the Digital Services Act’s restrictions on targeting, consumer protection rules on misleading marketing, and the EU AI Act’s transparency obligations for AI-generated content must be ensured. As regulations may vary across target markets, this requires keeping a close eye on the development of the global legislation and regulatory environment.”

Could interconnected AI agents managing different marketing platforms create new risks, particularly when their decisions conflict, reinforce errors or trigger unintended actions across campaigns?

“Badly built AI agents always represent a risk. We already know about content-related risks from algorithmic bias, hallucinations, and meaningless AI slop. The more complex AI-based marketing systems become, the greater the system-level risks: errors spreading from one agent or system to the next, feedback loops between agents that can multiply errors, and conflicting goals. Integrating the work of diverse agents always increases the system-level risk, which is why human oversight is usually needed.”

How will autonomous marketing reshape the responsibilities of CMOs, marketing professionals and technology teams, and what new skills will businesses need?

“The biggest impact will be on junior-level marketing professionals who manage direct customer touchpoints, produce marketing materials, and analyse market data. Their roles are likely to be gradually shifted to AI. CMOs and other senior-level professionals are impacted less, as strategic decisions will continue to require human reasoning and may even expand their role in determining long-term competitive advantage and balancing growth, profitability, brand, and risk.

“Interfunctional coordination across sales, marketing, research, and technology remains a key challenge, calling for multidisciplinary expertise. Data and AI literacy, as well as skills in analytics and prompting, increase their importance, while understanding the basic marketing logic, customer needs, and connections between marketing activities and their impacts remains salient. Critical thinking and the ability to actively question the outputs of analytics and AI are what will separate great marketers from the rest.”

Looking towards 2027, what practical steps should businesses take to introduce autonomous marketing responsibly, and how can they determine whether AI agents are delivering genuine commercial value rather than simply accelerating existing inefficiencies?

“Safeguarding consumer data privacy, managing algorithmic bias, and maintaining brand authenticity are among the most pressing challenges marketers face as they increasingly rely on autonomous AI decisions in marketing. The foundations of ethical and effective AI implementation lie in data quality and integration, as well as clear strategy, goals, and responsibility. Once a solid foundation is in place, AI should be introduced gradually, identifying which decisions to automate and which to keep in human hands.

“Ultimately, the commercial value of AI agents (like any novel tool) will be determined by their long-term impact on profitability and competitiveness in the marketplace. Efficiency (e.g., speed, accuracy, costs) should be compared with the decisions made without the help of AI. When competitors have access to the same agents, efficiency gains are often competed away; hence, they are also likely to erode over time. To stay alert to early warning signals from the markets, it is useful to keep a particularly close eye on developments in brand image and associations, as well as the firms’ customer relationships as intermediary measures of marketing performance.

“In my view, outsourcing more mundane marketing tasks to AI opens new opportunities for marketers to focus on strategic long-term planning, which too often gets left behind by short-term optimisation. Thus, AI agents may even strengthen marketing’s long-term role within firms.

“From the perspective of individual marketers, executing marketing campaigns has become table stakes. To set themselves apart from their peers, marketers cannot rely on mastering the latest AI tools. The best marketers will find ways to build on and expand what AI makes possible. Just as marketers need to consider the unique value propositions and competitive advantage of the offerings they sell, they will need to be equally concerned with the value they bring to the AI agentic world.

“Based on my yet unpublished research on the uses of AI in marketing, the benefits of AI are likely to vary across decision-making contexts; hence, identifying relevant boundary conditions for the use of AI agents remains a challenge and an intriguing topic for further research.”

Joh, Assistant Professor of Marketing at Aalto University School of Business.

Original source Johanna Frösén: Head-to-Head Interview

Back to home