| name | executive-operations |
| description | Diagnose operational constraints, process/vendor risk, execution capacity and automation opportunities, then turn them into measurable operating decisions. |
Executive Operations
Use this skill for process design, operational bottlenecks, vendor dependencies, service delivery, scaling, automation, cross-functional execution and operational metrics.
The default question is not “which tool should we buy?” It is what is constraining reliable throughput, why, and what change removes that constraint with acceptable risk?
Diagnose before prescribing
Map the current operating flow at the level needed to expose:
- demand/input;
- owners and handoffs;
- bottleneck/queue;
- systems/tools involved;
- manual workarounds;
- failure/retry paths;
- vendor dependencies;
- decision/approval delays;
- output/service-level expectation;
- observable metrics.
Classify the primary problem as one or more of:
- process;
- people/capacity;
- tooling/automation;
- policy/decision rights;
- vendor/supply dependency;
- information/measurement.
Do not automate a process whose purpose/ownership is still unclear unless automation is itself the controlled experiment.
Constraint-first operating rule
Throughput is often dominated by the tightest constraint. Prioritize removing the binding bottleneck before polishing low-impact steps elsewhere.
When a proposed improvement touches several areas, estimate the expected effect on:
- throughput / cycle time;
- error/rework rate;
- cost per unit/order/case;
- customer impact;
- human time reclaimed;
- reliability / failure blast radius;
- working capital or inventory;
- vendor lock-in/switching cost.
Process before headcount
Before adding people to repeatable work, ask whether the work can be:
- removed;
- standardized;
- simplified;
- batched;
- automated;
- moved to a clearer owner.
Headcount can be the right answer. It should not be the substitute for understanding a broken process.
Automation decisions
For automation, model:
- frequency and manual time;
- error rate/consequence;
- implementation/setup cost;
- maintenance/exception-handling cost;
- integration/security risk;
- expected utilization;
- payback and opportunity cost;
- observability and rollback/manual fallback.
Automate the stable path and explicitly design the exception path.
Vendor and dependency risk
For critical vendors/services inspect:
- SLA and actual performance;
- switching/export/migration path;
- data ownership/portability;
- renewal/termination terms;
- concentration/single-source risk;
- support escalation;
- security/compliance boundaries;
- failure fallback and recovery time.
Critical single-source dependencies should have a known switching or degradation plan even when a full second supplier is uneconomic.
Metrics
Pair lagging outcomes with leading indicators where useful.
Examples:
- revenue/output + qualified pipeline/order volume;
- missed SLA + queue age/capacity utilization;
- churn/refunds + defect/complaint rate;
- cost + process time/rework;
- deployment incidents + change-failure/rollback signal.
A metric that cannot trigger a decision is often dashboard decoration.
Decision method
- Define desired operating outcome.
- Map current flow and binding constraint.
- Identify root cause, not symptom.
- Compare process/people/tool/vendor options.
- Estimate impact, cost, risk and reversibility.
- Choose the smallest change that meaningfully improves the constraint.
- Assign owner and operating metric.
- Define rollback/fallback and review point.
Pair with executive-finance for capital/payback, executive-people for org/capacity, executive-legal for vendor/contract/regulatory risk, and engineering/release packs when actual technical implementation begins.
Failure modes
Avoid:
- buying software before diagnosing the process;
- optimizing non-bottlenecks;
- adding headcount to tribal/manual chaos;
- automation with no exception/fallback path;
- KPIs with no owner or action threshold;
- critical vendors with no exit knowledge;
- operational plans that ignore change management and real human workload.
Provenance
Original Agentit guidance materially informed by the operations specialist design and constraint/process heuristics in Sente Labs' OpenExecutive (Apache-2.0). See THIRD_PARTY_NOTICES.md.
