How holding companies are building AI capabilities amid rapid market change

Key takeaways

The promise and pressure of AI adoption in media buying

The conversation around AI within holding companies has matured significantly since 2024. What began as a technology topic has become a business strategy question. Executives at major media and creative holdings are no longer asking whether to adopt AI, but rather how to build sustainable capabilities that serve clients without creating new problems.

The Forrester and 4As collaboration study released earlier this year offers some quantifiable data to underpin these conversations. Amongst the findings: 74% of agencies use generative AI to summarise documents and communications; 70% apply it to research and competitive intelligence. Those numbers suggest that AI tools have moved past the experimentation phase and into operational workflows. Yet the same study reveals significant friction: 63% of agencies cite accuracy and bias as ongoing concerns; 62% raise legal challenges; 55% point to privacy and security risks.

For agencies that have made meaningful progress with AI integration, the narrative is less about the technology itself and more about the cultural and structural changes required to use it effectively at scale.

Building internal expertise whilst managing the learning curve

One of the clearest themes emerging from conversations with holdco technology leaders is the tension between the pace of AI change and the time required to build genuine organisational expertise. The learning curve is steep, and the executives responsible for leading AI initiatives are acutely aware that the success of their programmes depends on how well they communicate uncertainty internally.

As one technology executive put it: “It’s a lot of existential moments where you’re having to think, ‘How do I do this?’ and get an inner circle to help you figure that stuff out. Those learning curves can be tough because a lot of people are looking at you in your organisations and thinking you’re gonna have the answers.” That vulnerability in communication -being willing to say “I don’t know yet” rather than projecting false confidence -appears to be a critical factor in building trust as these programmes mature.

The holding companies that are progressing fastest with AI are those that have created protected spaces for experimentation, where failure is treated as learning rather than a loss. That approach requires both senior sponsorship and a clear articulation of what success looks like -which is itself a challenge, given how nascent many of these programmes remain.

The shift from specialist to orchestrator

The hiring conversation within holding companies has undergone a subtle but important shift. Rather than seeking specialists in AI or data science alone, major holdings are now seeking what they call “orchestrators” -individuals with deep expertise across creative, media, data, and technology disciplines, capable of making informed decisions quickly and translating between different functional areas.

The logic is straightforward: as AI becomes embedded across media buying, creative development, and campaign measurement, the traditional boundaries between functions become less relevant. A person who understands both creative brief-writing and media optimisation algorithms is more valuable than someone who specialises in one domain alone. The problem is that such people are rare, and the market for them is intensely competitive.

Holding companies are responding by investing in upskilling and training programmes, but the consensus is clear: finding or developing true orchestrators is one of the industry’s primary challenges. As one executive noted: “An orchestrator is a resource, an individual that has deep expertise in multiple areas. The orchestrator needs to be a master of everything. Where do you find those people?”

The governance bottleneck: clients can’t act as quickly as agencies can recommend

One of the most candid observations from the recent holding company roundtables concerns the growing gap between what agencies can identify as an opportunity and what clients can actually implement. Agencies, armed with AI-driven insights and optimisation tools, are generating more and better recommendations than ever before. But those recommendations often hit a wall on the client side.

The friction points are multiple: clients lack internal AI governance approvals; they don’t have creative assets ready to deploy; budget is committed elsewhere; internal stakeholder alignment takes longer than the opportunity window. As one agency executive put it: “The disheartening part is we get to a place where all of our tools can help identify better opportunities, and then you present it to your client and they can’t act on it. The client doesn’t benefit if we don’t help them untangle themselves.”

Some holding companies are working with clients to streamline approval processes by shifting from approving specific campaigns to approving strategic parameters -what they call “playing fields” of creative, media spend, and budget reallocation authority. If a client approves the framework in advance, the agency can act with agility within those guardrails. It’s a more sophisticated approach than traditional approval workflows, but it requires clients to have confidence in their agency partners and clarity about their strategic constraints.

Measurement and platform interoperability: the technical and political challenge

As AI creates new fragmentation in the media ecosystem -with SSPs becoming DSPs, direct integrations between advertisers and publishers, and more complex measurement challenges -the ability to measure how each channel and tactic contributes to outcomes becomes increasingly critical. Yet this is precisely where the ecosystem’s structural limitations become most apparent.

To measure contribution in a deduplicated way across channels, platforms need to integrate openly. That requires willingness to prioritise advertiser and agency needs over proprietary advantage -a tension that has existed in the media ecosystem for years. According to holding company executives, there is “a sea change in the openness” from platform providers, but the reality remains: some platforms still resist integration in ways that make end-to-end measurement nearly impossible.

This is not fundamentally a technical challenge. The tools and approaches exist. The bottleneck is commercial and political -some platforms benefit from opacity, and changing that requires either regulatory pressure or sufficient client demand to make openness a competitive advantage.

Speed and experimentation as core competitive advantages

For holding companies that have embraced AI, one unexpected advantage has emerged: the ability to experiment and test at a much faster cadence than was previously possible. Where traditional campaign development might take weeks to move from idea to deployment, AI-assisted workflows can compress that timeline significantly. That speed advantage only works if teams are comfortable with iterative failure -the willingness to “break some shit, and that’s OK,” as one executive put it.

That cultural shift -moving from “get it right the first time” to “test, learn, iterate” -is one of the most underestimated aspects of successful AI adoption. Holding companies that have made that shift report that their teams are more productive and their clients are seeing faster results. Those that haven’t made that cultural move are often struggling, regardless of their technology investments.

The business model question: AI as cost or as revenue opportunity

Amongst the most revealing findings from the Forrester and 4As study: 61% of agencies currently classify AI as a cost of business, whilst 31% plan to monetise it within the next 24 months. That shift -from cost management to revenue opportunity -represents a fundamental change in how holding companies are positioning themselves in their client relationships.

If AI becomes a capability that holding companies can charge for separately, or that allows them to deliver better results at the same price point, the economics of their business improve. But that requires being able to articulate clearly what the AI-driven advantage is and why clients should value it. In a market where clients are still grappling with their own AI integration, that conversation is not yet fully formed.

Takeaways for holding company executives and publishers

For holding companies building AI capabilities, the priority is clear: invest in people and culture as much as technology. Orchestrators, protected experimentation space, and streamlined approval processes matter more than having the most sophisticated AI models. For publishers, the implication is equally important: clients and their agencies are moving faster, and that speed will reshape how media buying works. Publishers that can enable that agility through platform integration and flexible capabilities will capture more value in the new model.

For more guidance on AI implementation in publishing and media, Publishrs offers a suite of tools that help publishers manage the complexity of modern media buying. Whether you’re building AI capabilities or optimising your platform integration strategy, the Publishrs team can provide strategic support. To understand how your publishing operation can better support the increasingly sophisticated demands of modern media buying, schedule a consultation with Publishrs.

FAQ

Source: Digiday