operational leadership
WHAT DOES AN AI-AUGMENTED OPERATIONS FUNCTION LOOK LIKE?
An operating function spends too much of its day reconstructing what is happening, repairing schedules and chasing records. The change that matters is that the live picture, the coordination rules and the expert methods become a shared operating layer. Judgment stays with the person. The operation stays current.
The schedule may be current.
The explanation is not.
Operations and delivery convert a commercial promise into completed work and cash. The accountable leader may be a COO, operations director, Head of Operational Excellence, country lead or business-area head. Around that leader sit resource managers, schedulers, programme managers, project controllers, quality specialists, supplier managers and site or regional managers.
The function drives growth through available capacity, reliable delivery and customer retention. It shapes contribution margin through utilisation, rework, subcontractor spend and the ability to turn completed work into an accepted deliverable. Unit heads often carry a local P&L beside the operating scorecard.
Eight processes- 01Capacity and resource planning
- 02Scheduling and dispatch
- 03Project and assignment delivery
- 04Quality control of delivered work
- 05Supplier and subcontractor management
- 06Site or unit follow-up
- 07Continuous improvement
- 08Incident handling and delivery reporting
The scorecard spans utilisation, billable share, gross margin, on-time delivery, schedule adherence, throughput, cycle time, backlog, rework, complaints, SLA attainment, incident frequency and forecast accuracy. The measures exist. The explanations arrive late.
Information sits across the ERP, field-service or project system, CRM, workforce system, quality records, Teams or email, customer portals and spreadsheets. The schedule can be current while skills data is stale. Finance sees invoiced delivery after the fact. The operational explanation remains in conversations and in the heads of experienced coordinators.
The function is rich in operating data and poor in shared operating context.
Units submit updates, central staff reconcile definitions, and management discusses what happened last month. Attention goes to the largest business area or the site with the loudest manager. Intervention starts after the assignment, schedule or customer promise has already moved.
Control becomes
continuous.
The function changes through two foundational shifts.
Management cadence moves from collecting reports to managing exceptions continuously. The operating record directs attention toward assignments, units and suppliers departing from plan. A deteriorating assignment becomes visible before its margin is gone. An implausible utilisation forecast is challenged before month-end. Small units receive the analytical depth previously reserved for the largest business area.
Monthly reporting remains useful for governance. It stops being the primary control mechanism. Operating meetings begin with what changed, what evidence supports it and which intervention now requires a decision.
Coordination and expert judgment become shared infrastructure. Scheduling rules, accepted quality examples, prior decisions and escalation thresholds enter the workflow. The best scheduler's constraints no longer live only in experience. The quality specialist's method no longer reaches only the work they have time to inspect.
Schedulers and coordinators remain essential, but the centre of their work moves. They manage exceptions, customer commitments, optimisation rules, supplier choices and cross-unit improvement. Their judgment reaches further because retrieval, comparison, documentation and monitoring happen around them.
Better coordination releases productive time around physical and professional work. It does not remove the technician visit, the regulated inspection or the difficult customer conversation.
Where it is weakestnovel trade-offs, customer commitments, safety decisions, regulated conclusions and work that depends on physical skill. These remain with qualified people, and should.
A day, a week,
a month.
A function that invests heavily still employs operators, coordinators, specialists and managers in 2028. Its shape changes. Less capacity sits in report collection, schedule manipulation and status chasing. More sits in exception management, customer recovery, data stewardship, workflow ownership and operational improvement.
The dayEach leader receives a role-specific briefing rather than beginning with a dashboard hunt. A field-service lead sees uncovered work orders, likely SLA breaches, late technicians, missing parts and proposed schedule changes. A professional-services head sees assignments likely to overrun, utilisation gaps, deliverables awaiting review and requested scope that no longer matches the plan. A multi-site leader sees only the units requiring attention, with the evidence behind each alert.
The scheduling layer recomputes options when work overruns, a customer cancels or a person becomes unavailable. It checks skills, territory, working hours, travel, parts and protected bookings. Routine changes inside agreed rules can proceed. Overtime, subcontractor calls, protected commitments and customer-sensitive moves go to a dispatcher for approval.
The delivery layer follows active assignments. It reconciles time, milestones, dependencies, customer correspondence and commercial scope. It prepares the next action and flags progress that conflicts with booked effort. Completed reports, certificates, analyses and work-order notes receive a first quality review against the method, contract, template and accepted examples. A named professional owns regulated conclusions, material advice and final approval.
The weekUnit heads receive a short operating review built from live records. It separates a data-quality problem from a delivery problem and shows what changed since the previous review. The meeting is spent choosing interventions, not arguing about whose spreadsheet is current.
The monthFinance and operations close the delivery picture together. Completed work, unbilled effort, milestone acceptance, subcontractor commitments and invoice readiness are reconciled before the management meeting. The formal report comes from the same operating record used throughout the month.
Continuous attention to every material assignment. A consistent first review of delivered work. Equal diagnostic depth for every unit. Capabilities once unaffordable at mid-market scale become the operating standard.
What runs, and what
stays with the person.
| Process | What the agent does | What stays with the person |
|---|---|---|
| Capacity and resource planning | Builds scenarios from backlog, role profiles and planning rules; with live access, maintains demand and capacity views by skill, site and week | Hiring, subcontracting and priority trade-offs |
| Scheduling and dispatch | Proposes the best qualified resource within travel, SLA, skills and working-time constraints; replans as conditions change | Exceptions, customer-sensitive moves, overtime and protected commitments |
| Project and assignment delivery | Maintains the assignment record, drafts updates, checks dependencies and flags scope, effort or margin drift | Interpretation, customer commitments and corrective action |
| Quality control of delivered work | Checks each deliverable against the method, contract, template and accepted examples; identifies omissions and unsupported claims | Regulated conclusions, material advice and final sign-off |
| Supplier and subcontractor management | Prepares work packages, checks qualifications, tracks responses and issues approved routine orders or reminders | Supplier choice, negotiation, exceptions and relationships |
| Site or unit follow-up | Produces a daily exception brief, compares relevant peers and records approved follow-up | Validating causes and deciding interventions |
| Continuous improvement | Mines delays, rework and incidents, groups recurring causes and drafts controlled experiments | Choosing the hypothesis, setting limits and verifying results |
| Incident handling and delivery reporting | Structures the report, retrieves similar cases, checks required actions and drafts internal or customer updates | Severity, safety decisions, root-cause judgment and external communication |
Four stages, and the
operations-system position today.
A realistic path has four stages. They map directly onto the rungs in The Access Ladder, and the order matters.
Exported schedules, SOPs, contracts, example deliverables and prior reports support planning scenarios, first quality reviews and drafted updates. Operators supply current context by hand. There is no live-system access and every output remains a draft.
Connect first to the system where delivery work is represented: the field-service platform, project tool, scheduling system or ERP production module. Add the ERP and CRM for financial and commercial truth. Workforce, quality, supplier and telemetry data follow. The result is one current view of work, resources and commitments.
The operating layer proposes schedule changes, task updates, acceptance requests, supplier orders and delivery-record changes. Responsible people see the source evidence and approve governed calls into the system. The record changes only after the decision owner has accepted the action.
Low-risk schedule moves, routine reminders, task creation and routing can run within agreed rules. Customer commitments, overtime, safety decisions, supplier awards and financial postings retain named approval. The boundary is set by consequence, not technical possibility.
Point connections are sufficient for one workflow. Cross-unit analysis across systems and years needs event-level history and shared definitions in a warehouse or lake.
The field-service, ERP or project system keeps permissions, transaction rules and the audit record. Each action records the initiating identity, source evidence, proposed change, approval, executed call and resulting system state. People stop opening several interfaces to assemble an answer. They receive the briefing, inspect the evidence and decide.
Six things we believe about
the continuous operation.
Operations is often treated as a tooling problem. It is a management design problem. Six beliefs determine whether the operating layer becomes part of the work or merely another interface around it.
Governance can stay monthly. Control cannot. A live record must direct attention to the assignment, unit or supplier leaving plan while intervention still matters.
Saving time on status collection is useful, but it is not the main prize. The deeper shift is giving every unit, assignment and deliverable the analytical attention once reserved for the largest problem.
The best scheduler's constraints, the quality specialist's accepted examples and the delivery manager's escalation thresholds should not remain private craft. Put them into the flow, keep named owners accountable and update them when operating reality proves them wrong.
No operating layer can warn early from stale schedules, late time records or disputed definitions. Updates should be proposed from actual work, approved where consequence demands it and written back so the next decision begins from the same truth.
Travel, equipment, human dexterity, customer conversations, regulated conclusions and safety calls remain real constraints. The operating layer prepares, checks, reconciles and monitors. Qualified people decide what consequence the evidence should carry.
Schedulers need to become exception managers. Coordinators need customer and control ownership. Quality specialists need responsibility for evaluation rules, difficult failures and accepted examples. If the old review work stays on top of the new flow, the operation gains another queue instead of a new way of working.
These six beliefs concern how work is delivered and controlled, rather than taking a position on technology. Every tool inherits the cadence, definitions and decision boundaries the leadership team chooses.
Roles, rhythm, and
where it fails.
Schedulers move from manipulating the plan toward managing exceptions and owning optimisation rules. Project coordinators move toward delivery control, customer follow-up and data stewardship. Quality specialists design evaluation rules, investigate difficult failures and maintain accepted examples. Site managers spend less time explaining results and more time correcting them.
The team needs process decomposition, data literacy, evaluation design, escalation judgment and a working understanding of system permissions. Senior operators must make tacit rules explicit. Junior employees need earlier exposure to difficult judgment cases because routine coordination provides less of their traditional apprenticeship.
The rhythmA document-grounded agent changes one coordination-intensive flow
A read-connected workflow establishes accuracy and escalation rules
Several units work from the same operating record
Roles, customer coverage, decision rights and the improvement cadence are redesigned
Where it failsUnreliable skills, duration or availability data. Every alert marked urgent. Definitions imposed centrally and disputed locally. Drafted updates resting on stale delivery records. Coordinators moved before exception ownership and customer coverage are redesigned. Generic language checks posing as quality review. Local workarounds continuing outside the governed flow.
Involve the people whose work changes while the rules are still being made. The operating layer becomes shared infrastructure only when operators trust its record, its boundaries and its path for escalation.