AI to make agencies lean, keen and driven by machine

Marketing agencies’ adoption of new technologies is enabling them to scale campaigns with data-driven, real-time precision but while automation will have huge consequences for the industry – from staffing levels and creativity to new business and agency remuneration – ultimately agencies will emerge leaner and smarter.

So says a new Forrester report, The Agency Workforce: Automation And AI Will Reshape Media, which claims these changes are already being seen, with more than one in ten (11%) agency roles – equivalent to thousands of jobs – likely to disappear as early as next year.

Adland luvvies may not like it but data science is now one of the key drivers of this new era. Data management platforms and agency audience activation technologies are already producing audience insights in a fraction of the time it takes to manually gather, process, analyse, and report qualitative and quantitative research. Insights that once took days and weeks are now revealed in hours.

Forrester cites Publicis Groupe’s Epsilon PeopleCloud platform, which registers more than 178 instances daily to reach 200 million consumers, providing near-instant understanding of consumer behaviour among audiences at an individual level.

Technology has also made time-intensive work within the studio production process faster and smarter.

For instance, Adobe incorporated its Sensei AI technology into Photoshop’s Lasso tool, speeding up image processing from hours to seconds, and Autodesk’s photogrammetry software allows creatives to take 2D photos and rapidly convert them into detailed 3D landscapes, the capture of which used to require days of set-up work.

Dentsu, meanwhile, uses machine learning to power its CGI studios which it is claimed has led to a 60% increase in the content production.

Technology is also revolutionising the decades-old agency services model that provides marketers and brands with human expertise to create, execute, and manage marketing.

Creative, digital, and media agencies are now wrestling with the possibilities and drawbacks of machine learning and robotic process automation.

MediaMonks founder Wesley ter Haar said: “You would expect agencies to be excited by automation and AI. But, ironically, an industry famous for innovation is holding back the tide.”

Despite lingering dissonance, agency holding companies have invested at least $12bn since 2014 in acquiring technology to power their martech, adtech, and data initiatives.

The melding of AI and intelligent automation (IA) with agency strategy and creativity has an impact on every aspect of the development of marketing communications. AI and IA can actually push the boundaries of creative craftsmanship. Software can now assist the intuitive, human process of creative development, resulting in a “human + machine” creative team.

One example is the martech application Persado, which uses natural language processing, a language database, and multivariant analysis to generate multiple campaign iterations and identify the best performers. The software helps determine effectiveness, such as selecting the word “nutritious” versus “nourishing” in advertising copy.

Matt McNeany, creator of Omnicom’s content platform ADZU, suggests that the volume of data and content used to train algorithms is becoming a differentiator. “Ads are better than [another agency’s ads] because the algorithms are better trained,” he said.

AI is also steering media planning and buying. While legacy systems and processes for buying broadcast media persist, all the major holding companies have built audience platforms to automate much of the media planning workstream.

Numerous agency audience platforms use AI to generate purchase journeys based on audience definition: Dentsu’s M1, GroupM’s mPlatform, Havas’s Converged, IPG’s Kinesso, Omnicom’s Omni, and Publicis’s Epsilon PeopleCloud all deliver AI-assisted planning.

Similarly, myriad programmatic adtech companies and agencies provide automated bidding exchanges to buy addressable media, including DV360, MediaMath, MightyHive, Seizmic, and The Trade Desk.

When it comes to staffing numbers, Forrester predicts innovations in AI and IA will ultimately lead to a long-term reduction in the number of employees by at least 25% by 2032.

This radical conclusion is grounded in dozens of interviews, an appraisal of the technologies that are already beginning to change how agencies work, data on the skills employed by agency jobs, and Forrester’s broader view on the future of work.

From content creation to production, broadcast buying to computer sciences, global client executives to HR and finance, AI and automation will transform every level of the agency on a task-by-task basis.

But Lynn Lewis, chief executive of UM in the States is not concerned about job cuts. She said: “Machines won’t take jobs. They’ll make them more interesting. They will take the highly monotonous and repetitive tasks, freeing up time for people to focus on higher-value, more impactful, and personally rewarding aspects of their work, resulting in more innovation and enhanced quality throughout the ecosystem.”

Moreover, Forrester predicts the rise of “intelligent creativity”, which is defines as: “A process of creative problem-solving in which teams of creators and strategists conceive, design, produce, and activate business solutions with the assistance of AI, intelligent automation, and data.”

In practice, this means AI does not replace agency creativity; it improves it. AI, intelligent software, and automation give marketers and agencies the ability to execute in volume using a deep understanding of audience interests instead of using intuition.

And Forrester argues that agencies and marketers must use data to inform decisions, not make them. Data is part of the answer, not the entire answer.

It states: “When CMOs and agencies let data make all the decisions, customers lose out. Data-driven display advertising and retargeting have created poor advertising experiences. According to Forrester’s data, only 40% of US online adults indicate that they find advertising a useful way to learn about new products and services. Instead, let data play its role in fueling intelligence and providing a base of knowledge on which analysts, strategists, creators, and marketers can build.”

Even so, by 2023, Forrester predicts that clients will be selecting agencies based on the best algorithm and data. The report adds: “Intelligent creativity will set off a software race among agencies for which can best wed software, data, and learning to their operational models. Machine learning isn’t the only key technology in play, but it suffuses all others, helping agencies complete tasks more effectively and aesthetically.

“Agencies with better data and better algorithms – though some will be commonly sourced – will benefit, as will those with more creative subject matter experts to partner with the machines. The result?

Marc Andreesen’s “software will eat the world” thesis will come to life as agencies differentiate on their software, data, and ability to reach algorithmic scale across clients.

Finally, the agency economic model will shift from paying for people to paying for platforms. Intelligent creativity will result in agencies being compensated for the AI and automation they offer.

This, Forrester insists, represents the first meaningful revamp of the current agency services/time compensation model in over three decades.

Agencies and clients will construct a hybrid model incorporating fees for software-as-a-service (SaaS) technology and advertising algorithms.

Digital agency 360i is already exploring the concept of a digital FTE. Chief media officer Doug Rozen concluded: “We’re starting to ask some fundamental questions about the value an agency brings by asking ourselves, ‘Do the bots become part of the staff plan?’”