In a world marked by dualities, AI represents both the greatest threat and opportunity.
Is AI just another buzzword or it can actually disrupt even the most creative segment of advertising?
In this study, we aim to cut through the noise of AI’s impact and potential, exploring how top advertising agencies from Romania are using, perceiving, and planning for AI.
Join us further to understand how to integrate these new tools not just to keep up, but to redefine what’s possible in advertising.
INTERPRETATION
Process & Cost optimization
In the dynamic realm of advertising, processes are often complex, but often characterized by meta- work (work about work) and/or excessive tasks that follow certain known paths.
Through this study, we’ve seen a consensus among participants to streamline processes that include repetitive or mundane tasks: typo corrections, pitch preparations, storyboards building, data analysis, SM research, content or design creation.
For any of the mentioned tasks, accurate automations are possible and implementations require meticulous prompt engineering mapped with the process for AI executing preliminary, and human oversight for more instructions & approvals to maintain quality and creativity.
Most of us are drowning in an abundance of tools, doing relatively the same tasks, using the same rules and knowledge, but in slightly different contexts. What if we can feed AIs with vast context and documentation of industry insights, research, guidelines, past executions?
Our experiments & estimates suggest that adopting such AI-driven workflows within advertising can increase productivity at least 25% leading to optimization of several FTs. This strategy not only promises greater operations but also increases employee retention from a burn-out stance.
Trends in AI
The AI landscape is ever-evolving with several trends emerging to redefine how industries, including advertising agencies, operate.
- AI Agents: AI agents are sophisticated software entities capable of performing tasks on behalf of their human users. Take into consideration each AI agent has defined roles, instructions, objectives, contextual memory, long-term memory or even tools they can interact
- Multi-agents: It extends by creating a decision-making chain of Imagine agents that prompt themselves where one’s output is the next one’s input. In the context of advertising, let’s take a complex example of a creative process normally involving AMs, Creatives, SM & more.
AI agents can manage tasks, from taking briefs, compiling the inputs to generate internal briefs for creatives, then copywriter AI agents can generate high-level ideas for the campaign, then SM agents decline the main ideas for multiple touch-points, and so on. - Few-Shot Prompting: This is a prompt engineering trend that relates to the AI’s ability to perform tasks effectively with minimal Few-shot prompting means training an AI on very few examples, yet still achieving high performance. Training your AI with past SM posts will help them align every new idea towards the referenced ones structure, tone, style & more.
- AI Image Generation: These tools can generate decent-quality images from textual descriptions, providing a quick turnaround on visual content and significantly lowering the barrier to entry for high- quality AI tools are capable of adapting to and reflecting brand books in their generated images. By analyzing previous content and preferences, some tools can customize their outputs to seamlessly match the brand’s specific style, ensuring consistency across all posts.
- Retrieval-Augmented Generation (RAG): One big issue AI have is context, and that’s what basically makes us great professionals, we combine great skills with knowledge (of a market, product, industry). The RAG leverages a smart database of data, it can retrieve data from a DB (e.g. brand archive) and augments the searching & generative process with contextual
- Agentic Scrapers: There are many scraping tools that can automate the process of extracting data from any For advertising agencies, such scrapers can be invaluable for gathering market insights, competitor analysis, and consumer feedback quickly and efficiently, especially when combined with agents.
New AI-enhanced scrapers are way more powerful because, at high level, they can simulate even browser controls, such as typing, clicking, scrolling, downloading documents, creating tabs.
Combined with a multi-modal AI that can understand images from websites, then it can decide the next steps and simulate real user behavior.
For example, you can build a social media research agent that extracts all data from influencer posts and, then, directly ask if there’s any collaborations with gambling companies? Same can be done with research agents to wide-web-browsing for pricing or competitive analysis.
Change in Business model
This is probably an unpopular belief, but I believe that the current business model of advertising agencies is broken by selling only hours. It’s not scalable, nor healthy. In our vision, AI should be covering mundane tasks with humans in the loop to validate & redirect outputs. Creatives should spend more to reach the best ideas, leaving AI again to craft the deliverables.
I believe that the future of agencies will shift from selling hours to selling solutions. Just like a software company, it can embody assets not only in people, but also in process-adapted AIs trained on the agencies’ proprietary data: previous creative campaigns or others.
Skepticism & Concerns
It’s still very early days in the commercialization of AI. That means it’s just as important to understand the concerns about AI’s trajectory as it is to understand how they currently use it.
AI Replacing Jobs: One of the more immediate and tangible fears is that AI will replace human jobs. In advertising, as AI can take over more, certain roles may become redundant. Agencies must consider how to reskill and reallocate talent to maintain job security and morale.
Oversaturation of Content: As AI tools make content production accessible and faster, there will inevitably be an increased volume across media channels. This challenge will shift towards
producing exceptionally creative, engaging, and unique content that can truly stand out. Agencies will need to push boundaries and innovate continuously to avoid being drowned out.
As AI increasingly integrates into advertising agencies, there is also a significant concern that the industry might experience a “McDonaldization” of creativity, where the drive for efficiency and consistency will overshadow the value of creativity.
Research methodology
For this study, we’ve mixed 15 to 20 both qualitative & quantitative questions to quantify AI adoption rates and effectiveness while also exploring deeper thoughts or concerns.
To get a representative & relevant mix, the respondents were solely executives, GMs and creative leaders from 80% of the advertising agencies members of UAPR. Of these leaders in advertising, 55% are from local agencies, while 45% are from advertising network agencies operating globally.
Mentions
The “State of AI in Romanian Advertising Agencies” was made possible by UAPR & WriteGPT.
Romanian Advertising Association (UAPR) is the professional association that brings together the leading agencies of advertising, media, digital and PR industry. Since 2000, UAPR has aimed to defend the interests of advertising agencies, develop and support the industry, and promote professional standards. UAPR is a member of the European Association of Communication Agencies (EACA).
WriteGPT is an AI productivity tool aggregating & personalizing most AI models, helping +20K users to write, adapt, summarize, and memorize content anywhere on the internet with the lowest friction. WriteGPT is part of Witloop, the tech-company helping companies develop custom AI- enhanced solutions based on their workflows & data.
