05.12.25
As we take stock amid the tumult of consolidations and mergers happening across our industry bandaged up as creating new stronger, leaner, enterprises, we shouldn’t lose sight of the fact that these are effectively defensive capacity reduction moves, which are our new reality as we move into the year ahead and beyond.
Let’s not beat about the bush, these moves are ultimately caused by a new technological shift in our midst that has compressed production times, improved personalisation and automated needs from creative to planning and buying, causing brands to query if they should renegotiate retainers and service agreements.
This isn’t a story about technology adoption. It’s a story about the impact on organisational redesign.
Yes we’ve had other technological shifts before, think of the .com bubble around the noughties changing work behaviours, the takeover of mobile just over ten years later (or even the industrial revolution or printing press if you want to go that far back!) – but the reality is we as humans and especially in our industry, have always been good at pivoting, finding a way to adapt and adjust our behaviours to best ride and make the most of the next crashing wave.
The questions for many right now will include what roles will be open to me tomorrow that I needed to start working on yesterday and which will be in most demand – that I can hone my invaluable experience gained (understanding people, storytelling and seeing patterns) towards.
These will include:
- Orchestrators (who blend human judgement with autonomous agent outputs)
- Exception Managers (who handle edge cases tech can’t solve)
- Lifecycle Stewards (who onboard, train, retrain, evaluate, retire AI agents)
- Human–AI Workflow Designers (who build processes that flex between autonomy & human intervention)
While beginning to conduct my own research to understand the impact to agencies & wider marketing communications, I’ve come across an interesting more general report from MIT Sloan Management Review that distils a more general business view called “The Emerging Agentic Enterprise: How leaders must navigate a new age of AI” which I’ve used as inspiration for this 5-minute read
Exec summary:
- AI breaks traditional organisational logic.
It behaves like both a tool and a worker, making existing structures, governance models, job roles, and investment frameworks inadequate – so they need to be redesigned
- Competitive advantage won’t come from early AI adoption.
Because AI is easy to access, differentiation will come from how organisations are redesigned around AI, not from who adopts it first.
- Org structures and roles must be redefined.
Expect flatter organisations, hybrid human–AI teams, fewer entry-level roles, and new roles such as AI orchestrators, exception managers and AI lifecycle stewards.
- Governance must become cross-functional and continuous.
AI can’t be owned by one function. IT, Tech, HR, Strategy, Legal, Finance and Operations must jointly manage autonomy, decision rights, training, oversight and risk.
- Investment and workforce planning shift to continuous lifecycle management.
AI systems need ongoing training, retraining and updating—just like employees. Workforce planning must consider humans and AI agents as part of the talent ecosystem.
Leaders today are facing one of the most destabilising shifts in organisational design since the arrival of the internet. Agentic AI systems that don’t just analyse but act which are forcing executives to confront a set of strategic questions that their current structures, investment models, and workforce plans are simply not built to answer.
What leaders are discovering is that AI introduces a new kind of contributor into the business: not a tool, not an employee, but something that behaves like both. And that challenges the very foundations of how companies operate, manage risk, allocate work, develop talent, and create value.
What Leaders Are Actually Struggling With
- “Is AI a tool, a worker, or something in between?”
Traditional management logic assumes technology either complements people or replaces them. Agentic AI does both—at the same time.
It can scale like a machine, adapt like a human, and learn like a team member.
This creates a conceptual and practical problem:
How do you manage a system that behaves like an asset in the budget, a worker in the workflow, and a strategic actor in the organisation?
Companies have no existing playbook for this.
- “The AI is spreading faster than we can redesign.”
Agentic AI is embedding itself into tools and platforms by default.
Leaders report that their organisations are already running on AI—even though they haven’t yet agreed on:
- Governance
- Decision rights
- Risk thresholds
- Workforce implications
- Team structures
- Career paths
AI adoption has become accidental, not strategic.
And this is creating a widening governance and capability gap.
- “We can’t rely on early adoption as a competitive advantage.”
Because agentic features are so easy to try, leaders can’t differentiate simply by “using AI first.”
Innovation theory explains why: AI meets all the conditions for rapid diffusion.
Which means competitive advantage shifts from who has AI to who can design the best organisation around it.
This reframes leadership’s responsibility:
The real battle is organisational design, not AI access.
- “Our structures, roles and workforce plans don’t match how AI works.”
Agentic AI cuts straight across traditional organisational boundaries:
IT owns the tech, HR owns talent, Legal oversees risk, and Business Units own workflows.
None of these departments can solve the AI problem alone.
Leaders are confronting questions such as:
- Who “manages” the AI?
- Who decides how autonomous it can be?
- Which roles shrink, evolve, or disappear?
- What new roles emerge?
- How do we reward employees when AI handles the routine work that used to drive promotions?
Most organisations are now discovering that their org charts assume a world without autonomous systems.
- “We don’t know when to retrofit or when to rebuild.”
Leaders feel stuck between:
- Retrofitting AI into existing workflows for short-term gains
- Reengineering the business around AI for long-term transformation
Retrofitting delivers quick wins—but caps the upside.
Reengineering offers step-change value—but risks being outdated before it’s finished.
This creates paralysis: the fear of doing too little vs. the fear of committing too early.
- “We don’t know how to supervise something that operates on its own.”
Leaders are used to choosing: human-in-the-loop or full automation.
Agentic AI breaks this binary.
It requires dynamic oversight that shifts by context, risk level and workflow.
Executives are asking:
- When do we trust the AI to act alone?
- How do we detect when it’s wrong?
- Who is accountable if it goes off track?
This represents a profound shift:
AI now requires management approaches previously used only for humans.
How AI Will Reshape Companies and Agency Structures
- Flatter, wider, more hybrid organisations
Agentic AI reduces the need for layers built around coordination, reporting, and routine analysis.
Managers will oversee hybrid teams of humans and AI agents, expanding spans of control.
Expect:
- Fewer layers
- More cross-functional teams
- More orchestration roles
- Fewer purely operational roles
- New roles emerge — orchestrators, supervisors, and AI lifecycle managers
Workforce planning will shift from hiring “operators” to hiring:
- AI Orchestrators (blend human judgement with autonomous agent outputs)
- Exception Managers (handle edge cases AI can’t solve)
- AI Lifecycle Stewards (onboard, train, retrain, evaluate, retire AI agents)
- Human–AI Workflow Designers (build processes that flex between autonomy and human intervention)
These roles don’t exist in most companies today—but they will become core.
- Dual career paths: augmented specialists and AI orchestrators
Career ladders based on performing routine analysis or coordination will collapse.
Instead, two paths will rise:
- Specialists augmented by AI (craft, deep expertise, domain judgement)
- Orchestrators managing hybrid teams (decision-making, ethical oversight, workflow design)
Agencies may shift from pyramids of junior labour to diamond-shaped structures with more mid-level generalists and fewer entry-level analysts which creates a question about how you maintain the talent tap!
- Processes become modular, flexible, and reconfigurable
Organisations will build workflows that can switch dynamically between:
- AI-driven efficiency
- Human-driven adaptability
This means designing for oscillation, not optimisation.
Agencies will adopt “embedded options” in workflows:
the ability to expand, contract, or reassign work between humans and agents in real time.
- Governance becomes a cross-functional, always-on capability
Traditional governance—once-a-year policy updates—won’t work.
Companies will need:
- Continuous, multi-department governance forums
- Context-specific autonomy thresholds
- Policy frameworks that evolve as fast as the AI learns
- Audit mechanisms modelled on how we manage human employees
Governance becomes a living capability, not a static function.
- Investment models shift from one-off projects to continuous reinvestment
Agentic AI learns, adapts and improves—if funded properly.
Executives will adopt investment strategies that resemble portfolio management, not tech depreciation.
AI becomes a learning asset, not a fixed asset.
This will reshape finance structures, budget cycles, and investment review processes.
The Leadership Challenge Ahead
The central question confronting every leader is no longer:
“How do we adopt AI?”
It is:
“How do we redesign our company or agency around a new class of organisational actor?”
Leaders must treat AI not as a technology project but as an organisational transformation that spans every function.
The companies and agencies that succeed will be the ones that:
- Recognise AI as both a tool and a colleague
- Build new structures rather than patching old ones
- Redesign work rather than automating tasks
- Upskill humans to supervise and orchestrate AI
- Fund continuous learning for people and agents
- Break down silos in favour of cross-functional governance
- Anchor everything in strategy—not hype
The winners will not be the fastest adopters, they will be the best designers.
Article by Christian Dam, Head of Agency.
