For a decade, enterprise marketing has been hooked on a quiet, expensive lie: that software purchase orders equal operational progress. Chief marketing officers assembled towering martech stacks to tame an explosion of digital channels, only to create an unmanageable performance tax. Gartner’s latest data confirms the fallout: martech stack utilisation has collapsed to 49%.
More than half of the software enterprise boards paid for sits idle, generating recurring subscription debt while army-sized agency teams waste thousands of hours manually stitching together incompatible data feeds.
Now, the advertising industry is repeating the exact same error with artificial intelligence.
The legacy drag
Over the past two years, legacy agency networks have treated generative AI as little more than a turbocharged copy-pasting tool. They have bolted standalone prompt interfaces onto fundamentally broken, multi-layered workflows, mistaking isolated task speed for structural transformation. McKinsey rightly calls it the “pilot trap”. Deploying fragmented applications atop brittle data foundations that fail to move the commercial needle by a single basis point.
This failure is not a technology problem; it is a business model conflict.
The advertising industry’s legacy holding companies are trapped in a classic innovator’s dilemma. Their business model remains stubbornly anchored to the billable hour. Selling human time creates a perverse economic incentive: the more friction, delay, and manual iteration built into campaign execution, the higher the fee.
Toxic efficiency
If an autonomous agentic network can generate, test, optimise, and deploy a global campaign in three minutes rather than three weeks, a time-billing agency suffers a 90% drop in top-line revenue. Efficiency is toxic to their P&L.
As a result, major agency networks promise “AI transformation” while quietly engineering ways to keep their headcount-heavy billing engines running. They offer clients automated efficiency while charging them for the hours it takes humans to oversee the automation.
The market will not tolerate this friction much longer.
AI stuckness
True enterprise resilience requires moving past superficial prompt engineering and embracing full agentic orchestration. The operational divide is already widening. According to Boston Consulting Group’s 2026 enterprise survey, while 96% of CMOs say they’re running AI initiatives, 42% remain stuck using AI as a basic administrative assistant.
Meanwhile, the top tier of market leaders has stopped experimenting. They are deploying autonomous, agentic systems directly into core commercial workflows—shifting from human-driven assembly lines to integrated intelligence layers.
This shift dismantles the traditional four-stage agency lifecycle:
- Strategy stops being a static quarterly presentation in a glossy deck. It becomes a live, predictive simulation continuously modelling consumer sentiment, pricing sensitivity, and competitive movement in real time.
- Creative production and activation merge. Instead of weeks of sequential asset adaptation and manual localisation, autonomous agentic systems dynamically construct, translate, and format creative assets within strict brand safety guardrails the moment a market opportunity emerges.
- Media optimisation moves from a post-campaign autopsy to live execution. Agencies have spent decades charging clients to analyse wasted ad spend after a campaign ends. Autonomous networks now reallocate capital across platforms and creative iterations mid-flight, long before media budgets burn out.
- Measurement finally abandons vanity proxy metrics. Powered by clean-room architecture and verified first-party data, performance is tied directly to bottom-line velocity, pricing power, and customer acquisition cost.
When campaign timelines collapse from months to seconds, hourly billing loses all commercial credibility. The dominant agency models of the next decade will be structured around performance equity, outcome-based risk sharing, and commercial growth arbitrage.
Critically, this restructuring will not be led by legacy holding company hubs in London or New York. High overheads, multi-year office leases, and public-market margin pressure make radical self-disruption nearly impossible for incumbents.
Instead, as I argued in The Brave Code, the most aggressive operational innovation is occurring across emerging markets. Unburdened by century-old billing habits, bloated middle-management structures, or heavy legacy infrastructure, agile agencies across Africa, Latin America, and Southeast Asia are leapfrogging the holding company era entirely.
They are building natively as intelligence engines—operating with lean, high-conviction strategic teams that govern autonomous operational networks.
The mandate for brand leaders and investors is stark. The era of building enterprise value through billable headcount and bloated software procurement is finished.
The agencies that survive will not be those selling time to manage software, but those that can turn real-time market signals into commercial value at the speed of modern culture.
