An Operational Forecast, Written as the Solution

A companion to An Operational Diagnosis, Written as a Solution, written the same way. Still not a vacancy. Where the first piece diagnosed how growing businesses fail now, this one forecasts the failures that arrive with AI adoption, and describes the role that would counter them.

The premise is simple. AI has collapsed the cost of execution. Work that needed a team now runs on one person and a model, and the scarcity businesses were built around, the capacity to do things, is disappearing. What replaces it is quieter and harder to see. The failures below do not throw errors. The business stays busy. It just stops being one system.

Director of Business Operations 

London Area, United Kingdom, D -24 Months

Actively considering opportunities 

As the business scales, we also continue to go through unprecedented transformation. With more of our work running fast, independent, and AI-augmented, we increasingly transition people into seeing across the whole and producing the ground truth leadership acts on. The role exists to counter the operational debt that builds quietly inside any growing business: the processes that no longer fit, the dependencies that go unmapped, and now the teams and systems that, left to their own loops, drift from the centre and from each other. Above the debt sits the harder line: keeping effort and investment pointed at where the business is going, not at whatever cheap execution now makes possible. The work is to hold the business coherent and answerable as people, systems and automation run with increasing independence.  

Success is defined by what the role produces: sensing, diagnosing, mitigating and automating where it scales, handling manually where it does not, to iteratively improve our systems. Each issue is recorded at every step, from the failure identified to the control built to catch it. The output is an implemented control, tested outside the system's own report and retired once superseded, each iteration building on the last. 

The role itself is new, and growing the people who can do it is part of the work. The experience that once built this kind of judgement from day one of a career is being automated away, so developing future leaders becomes deliberate rather than incidental. Without it, the default as those who currently provide judgement are not replaced is to hand more to the machine.  

The Role

A senior, cross-functional problem-solving role with a wide remit, reaching across systems and agents as much as people. The day to day varies; the work stays consistent. It carries the autonomy and delegated authority to hold the whole together and produce the ground truth leadership acts on. 

• Own business coherence. Ensure fast-moving, independent and augmented workstreams remain aligned to the central objectives of the business. Coordinate and consult cross-functionally, ensuring the business decision layer continues to be visible and viable. 
• Enable independent, self-directed augmented work. Build and improve business processes around core principles and standardised integrations, and where necessary develop localised modular systems that let each element move within its constraints. 
• Ensure reporting produces signal not noise. Define and refine the actionable metrics that measure success, identify failure, and prove progress, removing the rest from decision-makers. Set their format and interpretation. Own the periodic review that retires metrics as they lose value and surfaces their replacements. 
• Govern the decision boundary. Define and refine what humans decide and what is delegated to automation and agents, and the guardrails each works within. Set, test and review the exception criteria to intervene before the point of failure, not at it. Ensure all breaches are reviewed as part of the incident process. 
• Maintain independent assurance. Verify what these systems run on, decide and do, through testing and controls outside them, never their own report. Keep it independent of both the teams who build the systems and the owners who answer for them, with accountability resting on a person, not the machine. 
• Anticipate failure. Map where processes and independent, augmented workstreams will fail under scale or change, proactively stress-test, and lead projects to mitigate through innovation, transformation or iteration. 
• Direct incident diagnosis and response. Lead the structural diagnosis when things go wrong, including breaches flagged at the decision boundary, identifying root causes and facilitating the response. Coordinate with teams, agents and technology providers to understand the interplay between human-prompted and AI-interpreted decisions and actions. No issue is left unresolved or written off as an unexplained machine outcome. 
• Facilitate capability lifecycles. Bring new products, services, verticals, technology and agents into the architecture, reviewing each at inception against the current standard: absorb what fits, run outliers as modules, and where one is strong enough, make it the new standard. Retire cleanly what they supersede, so nothing runs unowned in the background. 
• Prioritisation and allocation. Work across the senior team and operating functions to align on priority, sequence, and trade-offs. Own the handoff points between functions and systems, and build alerts for inaction where flow stops or falls out of alignment, so it can be caught and fixed. 
• Develop emerging leaders Build the next layer of judgement, growing people who understand the business end to end and can govern its systems and the boundary with the machine 

Requirements 

• Demonstrate a track record of operating in a variety of environments and functions in leadership roles. Ideally with experience of designing processes in early stage start-ups and redesigning processes in large businesses.
• Evidence of working within complexity, prioritising against the wider strategy, what to implement and in what order, what to fix and what to retire, creating sense from chaos. 
• Experience designing and implementing operating models, processes, new products, or governance structures from scratch, or in turnaround or transformation environments.
• Comfort operating with high autonomy and ambiguity. While obvious issues will be raised, curiosity and a default to first principles will surface underlying issues and more complete diagnosis.
• Evidence of keeping humans answerable for automated and AI-driven work: making the call independent of the system's own report, interrogating machine reasoning as hard as human, and ensuring accountability sits with a clear owner rather than the machine. 
• Experience owning capabilities and agents across their full lifecycle, from integration through to clean retirement. 
• Evidence of separating signal from noise across every audience, from the board to the strategic client to the most junior teams, each needing different metrics drawn from near-endless data but the same clarity. 
• Strong communication across the business, from reporting to leadership and the board through to speaking with our core teams to get the ground truth and insights. Able to translate technical, operational, and commercial perspectives into shared language without losing substance.
• Track record of developing emerging leaders, building judgement through doing rather than instruction, and moving people from strong functional performance to understanding the whole and the principles behind business decisions. 

About the Future Candidates

AI has collapsed the cost of doing. Work that once needed a team and a process now runs on one person and a model, and the scarcity businesses were built around, the capacity to execute, is gone. What replaces it is a quieter problem. Autonomous work drifts from the centre and from each other, deference rises as fluent output goes unchallenged, accountability dissolves into the machine, and signal drowns in volume. None of it throws an error. The business is still busy, but it is no longer one system, and coherence, not throughput, becomes the scarce resource.

The person who fits is defined by ability rather than domain. They see across the whole system rather than down a single one, reading what it is doing rather than what it reports, and finding the structural failure before it surfaces. They hold coherence across fast, independent and augmented workstreams, designing the shared core and the modules that let each part move within its constraints without pulling from the centre.

They govern the line between human and machine: what is delegated, where the guardrails sit, which exceptions get caught before they fail. They keep accountability from dissolving into the machine, holding answerability with a person and verifying the work independently rather than trusting its own report. And where everyone else defers to a fluent answer, they interrogate it, the one check the machine cannot run on itself.

Underneath sits a different set of instincts. They subtract where everything else adds, sit with ambiguity rather than forcing it closed, carry authority through credibility rather than position, and build what holds after they have gone. Unusual together, these are the abilities the future turns on, and the ones the past had no place for.

What this is actually describing

The first piece argued that coordination is the function nobody owns. This one argues that AI makes that gap fatal rather than expensive. When execution was scarce, slow coordination was survivable because everything moved slowly together. When every team and agent can run fast and independently, drift compounds at the speed of the work. Deference rises because fluent output goes unchallenged. Accountability dissolves into the machine because no one can say where a decision was actually made. None of this registers as failure, which is exactly the problem.

So the scarce resource inverts. Throughput is cheap. Coherence is not. The role described above is the first piece's role rebuilt for that world: still owning the whole, but now governing the boundary between human and machine, verifying systems independently of their own reports, and deliberately building the judgement that entry-level work used to build by accident.

So what?

The test from the first piece still applies, with one addition. Alongside asking who owns each recurring issue, ask where in your business an AI-produced output was last seriously challenged. If you cannot remember one, deference has already set in.

The forecast here is not that AI fails. It is that AI succeeds, everywhere, locally, and the business quietly stops adding up. The companies that handle it will be the ones that treated coherence as a job before it became an emergency. That job needs an owner. Decide who, before the volume arrives.