From Back Office to Control Tower: How Creative Ops Wins the AI Commerce ‘Lottery’
Tamara Lover recently joined ICP, bringing her extensive experience and longstanding reputation in creative operations. Her appointment is a significant win for the company, strengthening ICP’s strategic capabilities, enhancing its ability to deliver innovative, high-quality creative solutions, and further establishing its leadership in the sector.
–
For marketers, the arrival of transactional capabilities within ChatGPT represents far more than just another digital touchpoint to manage.
According to Tamara Lover, global head of creative operations solutions at ICP, it marks a fundamental reordering of how digital business functions.
“I wouldn’t look at ChatGPT now as just another marketing channel. I would describe it as a new interface for commerce,” says Tamara.
This new interface creates a ‘seismic shift’ by erasing the traditional boundaries between browsing and buying. The linear journey from awareness to consideration to conversion is vanishing; in this new economy, the distance between finding a product and buying it has effectively disappeared.
“It completely collapses the purchase decision funnel,” Tamara explains. “Because right now you have, for the first time in conversational AI, the loop between discovery and purchase all happening in the same conversation.”
But this efficiency comes with a terrifying caveat for brands: the potential for invisibility. Unlike a traditional search engine results page (SERP) that might list dozens of ranked options, an AI interface offers extreme scarcity. If an agent only serves up a handful of recommendations, your brand has to be in the mix.
“Ultimately, whether your product gets shown or not is a lottery,” Tamara warns. “In terms of ChatGPT, let’s just say it’s only showing eight products, the chance that your product is going to be one of those eight is probably slim to none.”
In a world of algorithmic scarcity, traditional performance marketing tactics hit a wall. If you can’t game the list, you have to bypass it entirely.
Tamara argues that this technological leap necessitates a return to traditional marketing fundamentals to survive the ‘black box’ of AI selection.
“If your branding is so strong and so powerful, you’ll be asking for it by name,” she asserts. “I think in this new world where you have less control, where the conversion funnel has collapsed, the power of branding becomes even more valuable.”
Decoding the Black Box
While the collapse of the purchase funnel presents a clear destination, the algorithmic roadmap to get there remains largely opaque. Tamara compares the current state of AI commerce to the chaotic, experimental early era of search engine optimisation (SEO).
“It reminds me a little bit of the early days of SEO and trying to crack the code of how you could get your brand website to be as high up on the Google ranking as possible,” she says. “This is all so new that there’s so much we don’t know.”
Despite this uncertainty, marketers are not flying completely blind. Tamara notes that while the decision-making logic is a ‘black box’, the inputs are identifiable. She outlines three distinct data sources that brands must optimise for if they hope to be recommended by an AI agent.
“The only things that we truly know about what’s in the black box, there are three different data sources that will be used to respond and show different products to consumers,” she explains.
The first is the training data, which comprises the historical information ingested to teach the model. “That’s the foundational layer,” Tamara notes.
The second is live retrieval data. This is the mechanism that prevents the AI from being stuck in the past, allowing it to scan brand sites and articles for real-time context. Crucially, this relies on technical standards like structured web data and schema markup.
The third and final piece is commerce integration. This refers to the direct technical partnerships between the AI provider and transaction platforms. “That’s being connected with Shopify, Instacart, Klarna. These are the third pieces that it will pull information from,” Tamara says.
For brands, optimising these three data feeds is the digital equivalent of ensuring your shop is actually unlocked when a customer arrives. “The number one thing you want to do is make sure that you’re visible to these AI platforms so you could at least buy the ticket to be in the lottery,” Tamara advises.
Creative Operations as the ‘Control Tower’
Marketing leaders cannot control the AI algorithms, but they can control the ‘hygiene factors’ required to appear within them. This necessity has elevated a function previously relegated to the administrative shadows: creative operations.
“Creative operations is finally having its moment,” Tamara declares. “It has become the sexiest unsexy discipline within marketing.”
Historically viewed as the administrative “back office dealing in process and flow charts”, the discipline is undergoing a radical rebranding. It is no longer just about asset production; it is the strategic ‘control tower’ of the modern organisation.
“There has never been a more important time for creative operations,” she argues. “Because with the hygiene that we need to do in order to be successful in this new world, the people who are best placed to take care of those hygiene factors are creative operations.”
The role has fundamentally shifted. It is now about “orchestrating all of the experiences and the activities that are happening across multiple areas within marketing to enable the agility, the accuracy, the governance to take place.”
The Playbook: Operationalising for AI Commerce
To turn the ‘control tower’ concept into reality, Tamara outlines a specific operational playbook focused on three pillars: technology, data, and people.
First, the technology stack must be simplified and integrated. The days of siloed systems are over; brands must unify their creative assets with their product data. “The first thing that needs to happen is unifying creative and product data,” Tamara advises. “Connect your DAM and your PIM.” She warns against manual processes, advocating for “middleware or APIs” to ensure content flows automatically. “Every time you have to upload and download an asset, you risk losing metadata, you lose time, there’s risk for human error,” she notes.
This connectivity relies entirely on the second pillar: data. If an asset cannot be identified by the algorithm, it effectively does not exist.“Getting your taxonomy and metadata right will be an investment. But that’s gold,” Tamara says. She suggests that the single most high-value action a brand can take immediately is a ‘taxonomy/metadata workshop’. "Because if we can't find the assets and we can't track the performance, we're creating more black holes. There are enough things we don't know. So we want to make sure we have that visibility."
Finally, organisations must restructure their human talent. Tamara envisions hybrid teams where marketing, production, and technologists sit together.
This evolution will birth entirely new job titles necessary for the AI age, such as an ‘expert on governance’ to manage chain of custody, and ‘AI quality assurance’, ensuring human oversight remains in the loop.
“I don’t think we’re at the stage yet of headless production, everything just happening automatically without human interaction.” she says. “You will have people whose role is to do a QA at different points specialising around the AI.”
Machine Learning and the Brief
While the industry fixates on the flashy output of generative AI, Tamara argues that the real operational revolution lies upstream in machine learning. “It’s so funny when people even talk about automation as AI,” she notes, distinguishing between simple automation, asset generation, and the deeper strategic potential of learning models. “Machine learning is probably one of the most important things that brands could do.”
She points to a specific, chronic pain point where this technology can offer immediate value: the creative brief. “A great brief is worth its weight in gold,” Tamara says. “And one of the things that I see is just bad brief writing.”
Rather than relying on humans to fill out empty templates, operations leaders should be deploying large language models (LLMs) to ingest historical performance data and auto-populate the strategic direction, she suggests. “Using an LLM to help autopopulate the brief based off of really great insights, bringing in AI insights, data, as well as what has worked brilliantly in the past, is such a valuable bit.”
Throw the Baby Out with the Bathwater
For creative operations leaders, the mandate is clear: do not wait for permission.
Tamara’s call to arms is simple, “Lead the change. There is nobody better positioned in the organisation to lead the transformation than you.”
However, leading this change requires a willingness to abandon comfortable legacies. Tamara warns that incremental tweaks – ‘shifting pipes around’ – will not suffice in an AI-first world.
“The number one thing that I think we need to get into our minds is the way that we have done things will no longer serve us in the future,” she says. “And we need to be willing to throw the baby out with the bathwater. We need to be willing to start with a bit of a blank sheet of paper.” She compares the necessary transformation to construction work, “Sometimes you’ve got to level the house and rebuild.”
The immediate next step for any leader is to run a full diagnostic. “Audit first,” Tamara advises, urging leaders to identify bottlenecks across people, process, and technology immediately.
While the task is daunting, she emphasises that no one has the perfect map yet. The industry must move forward collectively. “Build a community around creative operations where we all learn this together,” she concludes. “We should learn from each other.”
While audits, metadata and taxonomies provide the essential ‘hygiene’ to enter the game, Tamara argues that the ultimate victory lies in something less tangible.
If the AI interface is a lottery that is increasingly hard to predict, the only fail-safe strategy is to stop relying on the algorithm to suggest you, and ensure the customer demands you. “Fundamentally, the exciting potential for change is a shift from a short-termism approach to more long-time, bigger branding,” she says. “I think it’s really going to cause marketers to think slightly differently about how to win in a world where you can’t necessarily control the things that you used to control.”
For creative operations leaders, the job is to build the control tower that handles the complexity – the data, the governance, the speed – so that the brand can get back to doing what actually wins the lottery: building an emotional connection that transcends the search bar.
ICP has just launched a new home for its Content Studio. Check out the portfolio, services and more here.
