Are in-house creative leaders right to think their teams won’t shrink as AI takes hold?

Russell is the ex-director of Oliver and is a specialist in building creative teams within clients’ organizations.

He is speaking with The Drum following the publication of a report by the In-House Agency Leaders Club, which looks at how creative teams are getting on with AI, where they’re using it and their perceptions of the technology.

It’s a robust piece of research, with around 50 in-house leaders taking part, but surprisingly only 22% of them thought AI adoption would result in their creative teams becoming smaller. Generally respondents were positive about the onset of AI, with 80% of them agreeing with the statement, “AI’s role in creativity will primarily be as an assistant to human creative teams.”

This year, the AI narrative seems to have moved from existential threat to practical benefits but where this leaves creative teams (and in-house creative teams in this case) is yet to become clear.

Earlier this year, Klarna proudly stated that it was cutting its external marketing spend by 25%, citing cost savings, increased efficiency and what it called “increased creativity”as benefits, and this is not an isolated example. Still, with its internal workforce being initially cut from 5,000 to 3,800, which will be wound down to 2,000, according to the BBC, it doesn’t look like this is the end of the process.

When you hear Russell’s view on the future of in-house creative teams in the round, it’s measured rather than pessimistic.

Can creative workers be redistributed?

Firstly, he argues that headcount may shrink, saying: “As AI that’s coming of age tends to be in optimization, production, distribution and delivery, redistributing that headcount to other areas is also going to be a challenge.”

This would affect the likes of art workers and those working with language believes Russell, who approaches the matter in a practical, dispassionate way and says: “I’d build my technology first with my process around it and then add the people, rather than build my technology around them. So I would always build from technology up, add process and then add the right types of skill sets around it afterwards.”

In areas where there may be less work, it is also an opportunity to retrain, Russell believes. “It’s not all doom and gloom and people losing jobs. It’s absolutely not. Good creative operators are retraining where they need to, as the tech will not take care of everything; it still needs to be properly facilitated.

“We just need to be grown up and realistic about creating better ways of working and utilizing technology at the best points.”

Building new teams around tech that improves productivity

Although this might sound blunt, Russell takes a pragmatic view as a typical brief for him from a client would be along the lines of, “Look at our technology ecosystem. Is it fit for purpose? How can we adapt and improve? Then we move on to the strategic and creative approach.”

The increase in generative AI capabilities for in-house creative teams is the perfect opportunity to reassess efficiency as it starts to illuminate bottlenecks, he reasons.

“When you stop and ask yourselves, ‘How are we operating at the moment?’ and realize you’re paying for four different asset management systems and three content management systems, you realize there are a lot of savings to be made. Firstly, you need a really solid operating model based on relevant technology.”

For Russell, Pencil is the most impactful piece of software for in-house teams. Its use case is that it can help practitioners make ads twice as fast while halving costs and doubling ad performance. So far it claims to have made more than 1 million ads across 5,000 brands.

“Pencil’s an unbelievable piece of technology, it’s proven and is delivering millions of pieces of content now,” says Russell. (It is owned by The Brand Tech Group, which also owns Oliver) and for Russell, “I think it’s the single best story of AI working in marketing today.”

What does Adoption really look like within an in-house team?

Hygiene, health and nutrition FMCG company Reckitt is on its own learning curve with AI, which includes some adoption and a lot of experimentation. Much of this is being led by Reckitt global design officer Lee Barnsley who makes the distinction that in-house creative team sizes are likely to be reduced “downstream” in activities such as “in-market activation within communications, in-store and digital, where the AI will inevitably be able to perform those tasks in the near future.”

Distinctive brand assets and “equity guardrails” will need to be in place so that AI can control brand activation, but when you get that right, “AI can happen on a much larger scale, deliver consistent results quickly… and tweak messaging to resonate with consumers in a more agile way,” he says.

Upstream, if a design team is focused on “visionary thinking and new brand platforms for global scalability” jobs are safer offers Barnsley as “designers are able to look at challenges across a broad range of issues, from consumer experience to social awareness and accessibility.

“Identifying and understanding a real human need through design-led thinking combined with empathy is something I doubt AI will be able to do anytime soon.”

He says these same design leaders will encourage “functional expertise” in AI built within in-house design teams to evolve their own business practices.

Proprietary AI and experimentation

Like Russell, rather than worrying too much about the existential threat argument, Barnsley says, “It’s crucial for every company to ask why they should adopt AI?” In Reckitt’s case, “we wanted to improve our business rather than replace [anyone].” This led Barnsley to commission a proprietary AI platform, using the marketing of Gaviscon and Finish products as pilot schemes.

According to Barnsley, the platform used custom-built GPTs trained on Reckitt data, “multi-modal” generative AI tools for content adaptation and localization, and an interactive interface to create, test, and refine new product concepts from consumer insights data.

“Early findings showed that by using GenAI, teams can reduce concept development time by up to 60%, while significantly improving quality,” he says, describing it as “a game-changer for new product development, with its ability to transform unstructured data into insightful new ideas and marketable concepts.”

Although Barnsley recognizes that team size changes could occur at other businesses, at Reckitt he says AI has “taken the drudgery out of repetitive tasks” with things like time spent on post-campaign media analysis being reduced by up to 90% but quality improving.

Chasing sustainability targets with AI

He’s happy to share further exploratory work too, which gives insight into where a multinational like Reckitt is heading with AI. The company has been using AI and generative AI in its bid to reduce its Scope 3 emissions by 50% by 2030, compared with its 2015 levels. (Scope 3 emissions are indirect greenhouse gas emissions that result from an organization’s activities outside of its operational control). Reckitt is working with its network of suppliers to achieve this.

Some 300,000 data points were collected to get insights into emission levels per product, using AI and generative AI to turn these into a faster and more accurate carbon footprint. This process would previously involve manual categorization, which was time-consuming and less precise, Barnsley argues.

In just under four months, Reckitt obtained precise emissions data for each of its 25,000 products, improving the accuracy of its emissions footprint by 75 times – moving from 333 representative products to 25,000, he says. This data revealed new ways to reduce emissions by 2030 which will significantly contribute towards Reckitt’s net zero target.

Where Barnsley is able to draw on case studies that show AI in action within his own in-house team, the benefits are obvious.

His view of product development at the more senior levels of in-house design teams is that it’s a helpful additive, unlikely to replace anyone anytime soon.

As product development can often involve multiple departments and stakeholders, at Reckitt the process is collaborative and often data-driven.

The near future

“AI can sort and analyze the data and give you opportunity to develop and iterate ideas much faster than ever before, but more importantly it removes subjective opinions around design,” says Barnsley, who thinks that this will be a powerful tool for designers.

He says his team is currently engaged in thinking about repetitive tasks that can be replaced by AI so they can focus on bigger projects that will deliver growth.

For Barnsley, further integration at in-house teams will lead to generative AI becoming part of a company’s digital asset management (DAM) portal – which is a central area to store media assets – allowing any employee to create content on-brand and for assets to be easily updated when a master brand is updated.

“It won’t be long before we see bespoke brand generative AI portals that can answer questions on a brand via a chat function or create content for a campaign based on a pre-determined style or tone of voice.

“It will even amend headlines and digital content in real-time in response to social listening and performance. And that for me is a very exciting future,” says Barnsley, but it remains to be seen what these kinds of developments mean for creative practitioners working at the coal face.