Built to withstand: navigating agency life in an era of AI and economic fire

The marketing agency sector faces what might be described as a “triple disruption”: economic pressure marked by inflation and tightening client budgets, an era of media abundance requiring new strategic approaches and the pervasive rise of artificial intelligence that might fundamentally rewire how agencies work. 

This isn’t a future threat. For agencies across the United Kingdom, this is the present reality, where the challenge lies not in predicting what might happen, but in responding thoughtfully to what is already unfolding. Agency by Agency co-founder Tom was invited to speak at Agency Hackers’ Built to Withstand agency leadership conference in London last week.

The economic context: pressure and opportunity in tension

The economic headwinds are real. UK GDP growth projections sit at 1.1% for 2025 and 1.2% for 2026, whilst average earnings are expected to reach 4.2% by the end of Q4 2025. These figures create cost pressures for agencies whilst simultaneously meaning that every pound of client spend faces intense scrutiny. A generation of marketing leaders with ten or more years’ experience has known little stability, limiting opportunities for long-term planning and contributing to marketing’s diminishing voice in many boardrooms.

Yet the sector continues to demonstrate resilience. The UK digital advertising market surpassed £35 billion last year, growing 13% against GDP growth of just 1.1%. In the first six months of 2025 alone, digital advertising revenue totalled £18.7 billion (IAB UK/MediaSense report). Within this, search accounts for 44% (£8.3 billion), digital video for 23% (£4.3 billion) and display for 16% (£2.9 billion, with 78% bought programmatically). Retail media, one of the fastest-growing channels, reached £1.5 billion in advertising spend in H1 2025.

What becomes clear from our own data from mapping 25,000 active agencies in the UK is that whilst the overall sector shows 8.1% growth, this masks considerable variation. Some agencies are capturing the available “oxygen in the room” whilst others face clients re-evaluating budgets and demanding smaller scopes. Subsector growth rates vary from 2.8% for design and branding agencies to 20.4% for Amazon/Marketplace specialists. The agencies thriving are often those offering measurable outcomes in growth channels while those providing more traditional brand-building services remain stable but with lower average growth when it comes to headcount and revenue.

The AI paradox: widespread adoption, selective disclosure

Perhaps the most striking disconnect in the current agency landscape concerns how agencies are using artificial intelligence versus how they discuss it publicly. Forrester estimates that 75% of agencies have in-house generative AI models, yet our own research shows that only 23% mention AI anywhere on their website.

And when AI does appear in agency communications, it’s mainly in subsectors like digital transformation and data analytics, where over half reference it. But in traditional creative categories such as design, media and video production, fewer than 20% mention AI at all.

This silence suggests a complex set of anxieties. Agencies may fear client reactions or worry about appearing to commoditise their expertise. There’s a real tension here: how do you maintain profit margins on time-and-materials billing models when AI drastically reduces the hours required for tasks? The hesitancy to discuss AI publicly whilst embedding it operationally hints at an industry grappling with fundamental questions about value, pricing and positioning.

Three phases of AI transformation

Matt McNeany, President at WPP Open, offers a useful framework for thinking about AI’s impact across three overlapping phases: streamlining and efficiency, expansion and opportunity, and future customer experiences (Co:Definery podcast, June 2025).

The streamlining phase focuses on making existing work faster and cheaper through automation. AI agents can handle data gathering, research, strategy development and action, whilst generative AI creates content including images, video, text and voices, adapts material for different platforms and monitors performance. This is where most agencies exploring AI currently focus their efforts.

The expansion phase recognises that new technologies don’t just automate existing work, they create new possibilities. History supports this. When Squarespace and Wix launched, development agencies worried, but these platforms grew the market rather than replacing professional services. The same pattern emerged with accountancy software and recruitment platforms. AI’s capacity to enable smaller teams to produce more diverse content, achieve deeper audience understanding and manage global campaigns suggests an era of expansion rather than contraction.

PwC’s 2025 Global AI Jobs Barometer, analysing close to a billion job advertisements from six continents, provides evidence for this optimistic view. Their research found that AI-exposed industries experience three times higher revenue growth per employee, wages rising twice as fast, falling degree requirements and gentler employment growth rather than sharp declines. The report suggests AI makes people more valuable, not less, even in highly automatable roles.

The third phase, that of future customer experiences, represents the most profound transformation. AI assistants from Apple, OpenAI and others will fundamentally alter how consumers interact with brands, potentially rendering traditional search and digital advertising obsolete. Websites may become real-time personalised experiences. The core pillars of marketing – search, media, content, CRM – face dramatic evolution or disappearance. This creates an enormous opportunity for agencies: helping clients navigate this entirely new landscape, defining brand roles without traditional advertising and addressing fundamental questions of positioning and differentiation.

The business model challenge

IPG CEO Philippe Krakowsky acknowledges that traditional time-based remuneration models have become unsustainable, necessitating a shift towards output or outcome-based pricing (The Drum, July 2025). This means charging for high-value deliverables and results rather than hours spent, forcing agencies to defend their value based on insight and transformation rather than task completion.

The tension is stark. If AI reduces the time required to produce work by 30%, 50% or more, agencies billing by the hour face proportional revenue decline unless they can charge higher hourly rates (difficult in a competitive market) or produce proportionally more work (challenging given client budget constraints). The mathematics simply doesn’t work in agencies’ favour.

Krakowsky notes that IPG now has “an avenue to generate technology and software revenue, not only remuneration for labour time” (The Drum, July 2025). Some large agencies are exploring Software as a Service models, selling their AI tools alongside traditional services. For smaller agencies without resources to develop proprietary technology, the path forward likely involves clearer articulation of value independent of time input.

Output-based pricing focuses on high-value units, such as ads optimised for a variety of channels, planning, strategic frameworks and creative concepts, rather than the hours required to produce them. Outcome-based pricing ties compensation to performance, though this remains challenging due to the many external factors affecting marketing results. The agencies most vulnerable are those still selling time; those most protected are selling insight, transformation and measurable outputs.

The two-track approach: optimisation and experimentation

Zoe Scaman’s critique of WPP’s AI strategy warns against over-optimisation at the expense of originality. She argues that friction fuels creativity, that more doesn’t equal better, that personalisation risks eroding cultural connection, that specialism matters and that pattern bias leads to safe, predictable work. Her central concern is that agencies focused purely on efficiency will sacrifice the very creativity clients need.

This leads to a crucial strategic framework: agencies should split their AI efforts into two distinct tracks. Seventy percent of resources should target incremental optimisation – automating routine processes, enhancing decision-making and improving customer experience for short-term ROI with timelines of one to three years. This is the necessary work of running an efficient business.

But Scaman argues that thirty percent must go toward radical experimentation, testing fundamental changes to business operations, exploring scenarios only possible with AI and accepting longer timelines of three to ten years with uncertain but potentially transformative outcomes, challenging core assumptions and actively shaping the future.

The agencies that will weather this disruption are those that can hold both approaches in tension: thoughtful optimisation of current operations alongside audacious experimentation with new possibilities.

What this means for talent and structure

One of the more contentious debates concerns junior hiring. A report from venture capital firm SignalFire found that AI has already led to a 25% decrease in entry-level hires from 2023 to 2024 at Meta, Microsoft and Google. Some businesses are responding to AI’s capabilities by reducing junior recruitment and increasing the billable capacity of senior staff.

Matt Garman, CEO of Amazon Web Services, dismissed this approach as “the dumbest thing I’ve ever heard,” asking: “How’s that going to work when ten years in the future you have no one that has learned anything?”

His point reflects a fundamental tension. Short-term, replacing junior work with AI-augmented senior staff might improve margins. Long-term, it hollows out the talent pipeline and deprives agencies of fresh perspectives. The agencies making this trade-off may find themselves in a decade’s time with an ageing workforce, no succession plan and an inability to attract the diverse thinking that drives innovation.

From the Agency by Agency perspective, this raises questions about what success looks like. Our Growth Report challenged whether growth should be the primary measure of agency success, suggesting that financial resilience, client relationship health, employee wellbeing and purpose-driven impact might offer more meaningful long-term indicators. 

An agency that prioritises profit extraction through AI efficiency gains over investment in future talent may show strong quarterly results whilst undermining its long-term sustainability.

Creating a compelling vacuum

Dhiraj Mukherjee, co-founder of Shazam, offers a phrase worth considering: “Great leadership is about creating a compelling vacuum”. In the context of AI disruption, this suggests that agency leaders shouldn’t merely react to technological change or economic pressure, but should actively include their teams and clients in the conversation about what agencies can and should be.

What can clients do in-house with AI, but still need agencies for? This question defines the compelling vacuum agencies must create. The answer likely involves:

Strategic guidance through uncertainty. Clients face the same AI disruption as agencies, often with less time to develop expertise. Agencies that can help clients navigate this landscape and not just implement AI tools, by fundamentally rethinking their approach to customer engagement will create genuine value.

Specialist depth in complex areas. Subsectors showing the strongest growth include those offering deep expertise in complex, rapidly evolving domains: Social purpose and sustainability (22% growth), Amazon/Marketplace (20.4% growth), Influencer marketing and Digital transformation. Clients increasingly pay premiums for agencies who can solve specific, complex problems exceptionally well rather than those claiming to “do everything.”

The human elements AI can’t replicate. Strategic thinking that challenges assumptions, creative work that connects emotionally, relationship building that creates trust, and the ability to synthesise disparate information into coherent narratives all remain distinctly human contributions. The agencies that will thrive are those that use AI to handle routine tasks whilst elevating human talent to focus on these higher-value activities.

Defining your agency’s path

Ultimately, agencies face clear choices. A narrow focus on AI for efficiency leads toward commodified expertise and talent loss. A broader embrace of AI for sustainable business and using it to liberate human talent for higher-value work whilst helping clients make sense of this new world, offers a path to resilience.

The agencies that will be “built to withstand” this particular inferno are those that:

Keep it simple. Focus on core values and capabilities rather than attempting to be everything to everyone. The trend across the sector is toward specialisation, with clients showing willingness to pay for proven depth in relevant areas.

Balance optimisation with experimentation. Improve current operations whilst simultaneously testing radical alternatives. Don’t let the urgent crowd out the important.

Rethink pricing models. Move toward output and outcome-based approaches that align your commercial success with client value rather than hours logged.

Invest in talent. Resist the temptation to hollow out junior ranks for short-term margin improvement. The cost of that decision may compound over time.

Make conscious choices about success. Growth isn’t the only measure of a successful agency. Financial resilience, client relationships, employee wellbeing, cultural health and positive impact all contribute to sustainable success.

This moment demands more than incremental adjustment. It requires agencies to examine fundamental assumptions about value, pricing, talent and purpose. The agencies that will thrive are those already past denial, actively engaging with the implications of AI and economic pressure, and making strategic choices that prioritise long-term sustainability over short-term optimisation.

The fire is real. But so too is the opportunity to build something genuinely resilient to withstand it.

FAQs

You will find information about all our data points within the report. For an overview of our methodology, the work with our partners at The Data City, and a glossary of definitions for all our data points, please take a look at our FAQs page.

Photo by Bethany Legg on Unsplash