Artificial Intelligence Agencies Agency Models

AKQA, Dept and Digitas on the race to launch AI creative services


By Sam Bradley, Senior Reporter

February 8, 2023 | 11 min read

In the nascent AI and machine learning services market, digital agencies are pursuing divergent paths on what to offer, how to build it – and how much they should charge clients.

A man and a woman looking at work on a laptop together

While some agencies have brought their AI service offers to market, others are still developing client propositions / Adobe Stock

In January, a red telephone appeared in the reception of London agency We Are Social. Guests that pick up the receiver (below) are invited to deliver a creative brief and are presented with a set of responses generated by the ChatGPT, which is linked to the phone by a Raspberry Pi rig.

Built by in-house creative technologist Pedro Garlaschi de Sousa, it’s the latest wheeze from an agency looking to educate clients on the potential of AI and machine-learning tools in marketing. Just last week, DDB EMEA unveiled a “fully automated creative agency powered by AI” which allowed users to submit a simple brief and receive a unique creative idea back. But elsewhere, agencies are swiftly moving on from an ad-hoc, experimental approach to AI services and towards formally presenting them as part of their product line.

Since the launch of ChatGPT, Midjourney and Stable Diffusion last year, agencies are scrambling to demonstrate their expertise in this nascent discipline, as well as explore potential cost-saving efficiencies within their own businesses. This week alone S4 Capital’s Media.Monks revealed that it plans to bring such a service line to market in summer while Ogilvy Paris announced ’AI.Lab’, a division which will sit within the group's global emerging experience unit. But these aren’t the only agencies racing to own the story.

A red telephone on a reception desk

Artificial strategy

Some, like Dutch agency Dept, even enjoy something of a headstart on the competition. Isabel Perry, vice-president of web3, says it’s been helping clients build up infrastructure to run AI and machine-learning models to improve their data capabilities for the last five years. “We build them the foundation, then pick up and run with the actual data science side,” she explains.

“We partner with clients to engineer AI models, either with some of Google’s own AI models for advertising solutions, or building custom solutions on top of a suite of different ML services such as Azure and Google Marketing Cloud,” she explains. Its data science services are delivered through a proprietary “Swiss Army knife” platform, Ada, that the company has offered for four years.

“It’s been built with the view that everyone should be able to use AI technology with brilliant results – like reducing the cost of creating content or increasing the speed of identifying priority products. It’s very powerful.”

Elsewhere, Publicis digital agency Digitas, has its own ‘AI Lab’ in place to identify AI tools that might be useful for the wider business. It’s created several products, including EmotionXD2, a natural language program that analyses masses of customer reviews and online comments to tell marketers what their customers are thinking about a given brand. Data science partner Louis Vainqueur tells The Drum that “we use this to map every part of our clients’ customer journey to identify gaps and opportunities.

“For example, we can glean insights such as customers expressing anger towards a return policy which we can then use to inform the experience.”

It also offers a service called Incrementality, which models digital media ROI for clients. Vainqueur says the agency aims to “translate the fantastic progress achieved in the AI community in the last five years into better-performing models that give a competitive edge to our clients.”

Over in the US, Boathouse – a Massachusetts-based marketing consultancy that caters to the healthcare sector – has spent the last two years working with a Nobel Prize-winning economist to develop a platform it calls Narrative Transformation.

The platform is designed to take in a morass of data points from different sources and present analysis and insights on a dashboard for C-suite decision-makers. In simple terms, it helps to measure and present ROI for digital transformation projects.

Founder John Connors says that it offers “incredibly relevant data on the company, employee, and CEO narrative and how that holds up against the company's transformation strategy. The result is that all the downstream tactics are far more informed, targeted, and impactful and the CMO is once again a driver in the boardroom.”

Though only formally launched in January, the platform is already in use with three of Boathouse’s clients. Connors says he expects it to become “the future of the agency and our core strategic driver.

Suggested newsletters for you

Daily Briefing


Catch up on the most important stories of the day, curated by our editorial team.

Ads of the Week


See the best ads of the last week - all in one place.

The Drum Insider

Once a month

Learn how to pitch to our editors and get published on The Drum.

Generating creative

Strategy represents a key application for AI tools in business. But while Boathouse’s effort is focused on reading and analyzing data, it’s the use of generative AI applications in the creative and content sides of the business that has garnered the most interest.

It’s also likely to be cheaper. The infrastructure and commitment required for large-scale data analysis mean that products aimed at aiding marketing strategy come at a premium. “It’s a really big undertaking. It’s very involved,” says Perry. In contrast, because generative AI can bring savings and scale to the creative process, it’s generally being pitched to clients as a middle-market solution.

Perry says that Dept has been integrating ChatGPT and Midjourney into more of its creative work lately, a practice likely to expand. “It’s completely swept through the company. People have woken up to the productivity impact,” she says.

The benefits for smaller agencies are apparent. Italian production agency Mono-grid began integrating AI and machine learning applications into its creative toolbox back in 2019 and recently began using OpenAI text and image tools in its general practice. Founder David Hartono tells The Drum that “we began using AI/ML tools for clients by utilizing deepfake technology to recreate a famous deceased painter. Since then, we’ve expanded our offerings to include other techniques.

“We estimate that 20-25% of our projects utilize AI/ML directly. Another 40% of our projects use AI indirectly in the research and creative process. Our experience with this integration has been positive and we are optimistic that this trend will continue to grow.”

At AKQA, staff have been toying with the applications of machine learning and AI for “the better part of a decade,” according to Geoff Northcott, chief experience officer. Most of its recent use cases involve souping up clients’ transformation efforts – using data to profile customers when they request assistance from customer support, for example, or building out digital product selection tools for L’Oréal.

“The demand from clients all of a sudden has gone through the roof. We’ve got about four or five clear use cases we’re building our offerings around because there are clear needs and client benefits.”

He says the agency is currently in the process of “codifying” its AI offering to clients. According to Jonathan Bolden, executive director of transformation, AKQA is initially focusing on marrying its customer experience applications with data and modeling uses, in line with broader trends in marketing investment.

“We are across more and more non-advertising investments, in apps, in websites, in stores. Our clients want to know which of these are the best investments… if we could take that data and use AI and machine learning, we could create a model like that smart enough to tell us what the best investments were across your entire experience,” he says.

The agency is farther away from bundling up discrete services around creative uses for generative AI. Currently, the focus is “less creative generation, more creative analysis,” explains Mike Procelli, executive director of data and analytics. He adds that the same expertise applied to media and CX spending analysis is being used to test creative assets at high speed.

“Theoretically, we save time and investments in A/B testing, because these tools can tell us which experiences or creative are ultimately going to win. It gets us to the answers that we’re looking for much quicker.”

Bolden says the end goal will be to use AI applications to bridge the data, strategy and creative disciplines. “We’re developing a growth model for a global CPG to help them target and acquire new audiences. I could imagine that at their scale, it’d be really easy to plug in creative inputs and generative AI against that growth model. I think that’s not far off.” Ideally, he concludes, “the power will be pairing those two things together.”

”We see those two things going hand in hand down the road, and it will really help our clients gain a lot of speed and a lot of efficiency.”

Artificial Intelligence Agencies Agency Models

More from Artificial Intelligence

View all


Industry insights

View all
Add your own content +