Last year I spent time at an underground site that had built activity dispatch into the bones of its Mine Control Centre. I went in sceptical. For years I had watched underground FMS sales pitches talking up activity-based planning without ever really buying into it.
The objections were obvious. Building the activity plan looked like serious overhead with no owner. Keeping it live when the first breakdown hit at 07:40 looked like a full-time job nobody had. Dispatch would already be at capacity. The whole exercise seemed to assume mining was more predictable than it actually is. It also assumed people existed to run it.
Mid-shift, the plan came apart. The dispatcher took the call and waved the Control Room supervisor over. That supervisor had total authority underground. On a conventional site, that response is a radio call: get the supervisor on the radio, wait for them to pull over, brief them, explain what equipment is where and what it is doing. The Control Room supervisor here had already read the impact on the daily plan, weighed the options, and redirected the fleet before that first call would have finished. No second call. No “let me check with the boss”. The decision happened in the room with full understanding of the impact to the plan.
I get it now.
Open-pit operations have not been seriously exposed to this concept like underground has. Maybe that is because sequencing is paramount to achieving the plan in the underground. In open pit, sure, it is important, but it will not bite you as hard.
That ends within five years. Activity-based dispatch becomes the default for open-pit mining, and Mine Control runs the whole pit instead of just the haul fleet.
The mechanism is already shaping. Caterpillar acquired RPMGlobal earlier this year. Hexagon’s OP Pro and UG Pro are ingesting activities from MinePlan natively. Modular has launched its open-API Ecosystem and bought Octodots for execution-layer AI. The harder question is whether the operations buying the software are ready for the operating model change underneath it.
What underground actually got right
The thing that converted me underground was watching a single plan flow from the planner to the dispatcher to the operator without translation losses. The dispatcher’s screen showed the same activities the planner had built. The operator’s tablet showed the same activity the dispatcher had assigned, with the expected duration and the productivity required. When an activity was at risk, everybody saw it at the same time. There was a single source of truth for what the shift was meant to deliver, and the system enforced it.
That is not how it works on most open-pit sites today. The 24 hour plan lives in a spreadsheet, set each afternoon for the next 24. The dispatcher gets a hand-off at shift start. The operator gets a verbal instruction over the radio. Three different versions of the same plan are running through three different heads, and none of them are visible to each other.
The systems doing this well are mature. Pitram has been running shift planning across underground sites for years. Sandvik’s Deswik Ops links Deswik schedules into the Deswik Operator tablet at the face. Mobilaris ties shift status to spatial plan compliance and surfaces it in a 3D control room view. Hexagon UG Pro pulls activities straight from MinePlan and reports field actuals back to update the schedule on the fly. The capability is mature, deployed, and operational. It has been confined to underground because that is where the constraints forced it.
What activity-based dispatch actually delivers underground is four separate things, each one valuable on its own:
- A systemised plan that gives the same direction to the dispatcher and the operator, in the same words, at the same time
- Spatial Short Interval Control measured at the minute, not at the shift
- Real governance around plan changes, not informal radio amendments that nobody captures
- Dispatch anticipating production changes before they bite, instead of reacting once the queue has already built
Each of those translates cleanly to surface operations. The reason open pit has not adopted them is not that they do not work above ground. It is that nothing has forced the change. Several things are now forcing it at once.
What open-pit FMS does not know today
Take any open-pit FMS today. The system knows which truck is loaded with which excavator. It knows the cycle time, the payload, the dump location. It knows every state code and every delay event. The data is dense and the dashboards are good.
What the FMS does not know is what the loading unit was meant to be doing and where it is meant to be next.
The plan for today’s shift was built in a spreadsheet by the planning and Ops team the previous afternoon. It has digger priorities, dig locations, active tip heads, ore targets and grades, cleanups, pattern prep and other miscellaneous tasks that are critical to achieving the plan.
Everything outside the load-and-haul cycle sits in the same gap. Drill pattern preparation for the next blast. Dump construction sequencing. Ramp work tied to a specific bench advance. Drill rig moves between patterns. Blast windows. Dozer and grader scheduling on the haul roads. None of it lives in the FMS. All of it lives in spreadsheets, whiteboards, and the supervisor’s head, with radio calls bridging the gaps.
This is the operational reality the FMS vendors are now moving on. It is also the gap that activity-based dispatch is designed to close.
Dynamic inside activities, deterministic across them
The strongest argument against plan-driven open-pit dispatch is not vendor inertia. It is that the surface environment is unpredictable. Open-pit operations get hit by weather, road damage, wall movement, and breakdowns at a scale underground does not face. A rigid plan published at 06:00 is dead by 08:00. Dynamic dispatch has earned its place because it is the only model that survives a real shift.
That argument is correct, and it is also incomplete. Activity-based dispatch does not require dispatch to stop being dynamic. It requires dispatch to be dynamic at the right level.
Inside an activity, dispatch stays dynamic. When a digger is on a face cleanup and the queue starts building, the dispatcher cross-allocates trucks to the next unit with capacity. When a truck is delayed at the fuel bay, the algorithm reshuffles assignments to keep the diggers fed. When a tip head congests, the system reroutes. None of that changes. It is the recovery role dispatch already plays, and the truck-and-shovel optimisation engine that has been the core of open-pit FMS for thirty years stays exactly where it is.
What changes is what sits above it. Across activities, the system becomes deterministic. The sequence is the sequence. Activity B7 is done before B8 starts, because the sequence dependencies say so. The blend is the blend, because the mill has committed to it. The drill rig move happens at 14:00 because the ramp will be clear and the drill rigs will not create an event. Deviations from the planned sequence are flagged events, not silent drift.
Most sites today run pure dynamic dispatch with no upper layer. Activities get reordered on the fly because the spreadsheet plan does not have any teeth. The result looks flexible but is actually undisciplined. Trucks are optimised. Individual cycles are optimised. The operation is not executing a coherent plan and therefore not optimised. It is reacting to the morning brief. The hybrid model is more disciplined than what most sites run today, and more resilient than the rigid factory plan the stochasticity argument is actually targeting.
Why open-pit FMS vendors are acquiring mine planning software
The structural signals are clear. Caterpillar acquired RPMGlobal earlier this year. RPMGlobal owns XECUTE, the live shift-level execution planning tool that builds optimal short-term schedules and feeds them into FMS. The acquisition does not make sense unless Cat intends MineStar to ingest XECUTE plans natively, push field actuals back, and run conformance against them. That is the entire point of paying that kind of money for a planning vendor.
Hexagon has been quieter commercially, but OP Pro and UG Pro ingest activity-based schedules from the MinePlan Activity Scheduler today. The architecture for the full closed loop sits inside the suite, and it is shipping. Verified named production deployments running the full bidirectional loop on surface are still thin in the public record, but Hexagon is the only one of the four vendors with the loop actually closed.
Cat is the inverse. After RPMGlobal, Cat owns both ends but has not connected them. There is no public statement that XECUTE plans currently ingest into MineStar Fleet, no integration timeline, no named production site running the loop. The acquisition is the bet. The architecture is not yet built.
Modular Mining took the other road: open APIs, customer-choice planner, no native planning-suite acquisition. Same destination in theory, different stack. Modular has put capital at the execution boundary. The Modular Ecosystem launched in September 2024, Octodots was acquired the same month for execution-layer AI, Applied Intuition followed twelve months later, and DISPATCH Look Ahead is shipping.
The wiring story is real. The data-model story is not. None of those moves announces a destination inside DISPATCH for activity-sequenced plans, sequence dependencies, or conformance reporting. Modular has the API surface. Whether there is anywhere to put what comes through it is an open question.
Wenco has neither. No acquisition, no internal development, no roadmap statement at the planning-to-execution layer, and no data model inside Wencomine for activity-sequenced plans to land in even if a custom integration tried. Wenco markets a feature called “Activity Dispatch” inside Wencomine, but the naming is misleading: the feature is truck-to-activity assignment, not the planner-to-FMS layer this post is about.
The absence is corporate-level: Hitachi rebrands to LANDCROS Corporation in April 2027, and not one LANDCROS-branded bet touches the planning-execution boundary.
Time will tell, but the absence reads more like a gap than a hedge.
The operational forcing functions are doing the rest. Autonomous Haulage Systems do not tolerate ad-hoc dynamic dispatch. They need deterministic spatial plans because the safe-interaction logic requires it. Mills are demanding tighter blend control, which means dig block level execution, not rough cuts. Per-tonne emissions reporting falls apart if the FMS does not know which activity each cycle was serving. Each of those is a separate pressure, and they are all pushing the same way.
This is not a vendor narrative. It is a structural reorganisation of where the planning-execution boundary sits.
What activity-based open-pit dispatch looks like in five years
Inside five years, the open-pit FMS does not look like today’s load-and-haul matching engine. Every planned task is a tracked activity in the system. Loading. Hauling. Drilling. Pattern prep. Dump construction. Ramp work. Blast clearance. Blend windows. Dozer and grader scheduling. The spreadsheet daily plan is gone.
The planner has full visibility of the plan, the constraints, the critical path and conformance from wherever they sit. They see what is on track, what is at risk, what has slipped, and why. They are not in Mine Control. They do not need to be. The system shows them the picture in real time, and the data flowing back lets them refine tomorrow’s plan based on today’s actual.
Execution lives in Dispatch. The dispatcher optimises trucks inside activities, runs SIC across activities, escalates exceptions when an activity is at risk, and is the only function with the live picture and the authority to act on it. Mine Control becomes the operational nerve centre for the entire pit, not just the haul fleet. The dispatcher’s screen is the single source of truth for what is happening on every bench, with every loading unit, every drill, every ancillary piece.
The resource consequence is structural. Dispatch absorbs scope that today is shared across the planner, the supervisor, and the radio. The cognitive ceiling is real, and the Underinvestment problem gets worse, not better, as more activities flow into the system. More activities tracked, more SIC, more exceptions, more decisions per hour. The headcount, the FMS specialist support, and the room design all have to scale with the scope, or the model collapses inside a quarter.
This is the part most sites will get wrong.
What sites will fumble
The technology will land. The vendors will ship (with varying levels of success). The harder question is whether the operating model around the technology is ready. Most sites are not, and they will not get there by accident. Five failure modes are predictable enough to name now.
| Failure Mode | Why It Happens | What It Looks Like |
|---|---|---|
| Authority for Mine Control beyond truck allocation | Field-based supervision is the cultural default. Dispatch authority today stops at the haul fleet. Extending it to drilling, dumps, ancillary equipment, and grade control is a power and accountability shift, not a software change. | Dispatch sees the activity at risk but has no authority to act on it. The supervisor in the field is still the only one who can move a dozer or hold a drill. Plan compliance becomes the planner’s report and Dispatch’s frustration. |
| FMS vendor capability divergence | Some vendors will deliver native activity ingest, conformance reporting, and SIC. Some will retrofit a thin layer on top of a dispatch engine. The differences will be invisible in the demo and visible in production. | A site signs a five-year contract on the strength of a polished demo and discovers eighteen months in that conformance reporting is rebuilt every quarter from exports. |
| Insufficient resourcing of Dispatch | The expanded scope assumes Dispatch can absorb more activities, more SIC, and more exceptions on existing headcount. The cognitive ceiling is already at the limit. | Dispatchers triage the new workload by ignoring half of it. SIC happens for load and haul. Drill, dump, and ancillary activities are tracked on paper or not at all. |
| Daily plan development and entry into the FMS | Nobody owns the pipeline from short-term plan to FMS today. Building activities, pushing them in, and keeping them current as the shift moves is new work. The planner thinks it is operations. Operations thinks it is planning. | The plan goes in at 06:00 and is dead by 09:00. By mid-shift Dispatch is running on radio calls again. The system reports conformance against a plan that no longer exists. |
| Management Operating System redesign | Activity-based execution alters meeting cadence, KPI ownership, decision rights, and accountability flows. Most sites will buy the software and skip the operating model redesign. | The system goes live, the dashboards work, the daily start-up meeting is the same as last year. Within six months the technology is treated as overhead and informal practice fills the gap. |
Each one has the same shape. The vendor lands the capability. The site does not change the operating model around it. The investment underdelivers, and the conclusion drawn is that the software did not work, rather than that the operation did not change.
The Management Operating System change is the one that will hurt the most. Mine sites have well-developed cadences for production reviews, weekly planning meetings, monthly performance reports. None of those are designed for an environment where the plan is alive on a screen, conformance is real-time, and exceptions are escalating from Dispatch to the planner during the shift. Without redesigning the cadence, the new architecture gets layered on top of the old one and adds cost without changing behaviour.
The Daily Plan ownership question sounds smaller, but it is just as load-bearing. Today the short-term planner ends their work when the spreadsheet is sent. With activity-based dispatch, somebody has to translate that plan into sequenced activities in the FMS, push it into execution, and amend it in real time when the first breakdown hits. That person does not exist on most sites’ org charts. If the site cannot answer who owns this work before the FMS goes live, the system goes live without an owner, and the plan is dead before lunch.
Why the operating model matters more than the software
The vendors will land some version of the execution layer over the next five years. Hexagon is the only one shipping the closed loop today, and even that on a thin set of named production sites. Cat owns the planning IP after RPMGlobal but no integration timeline. Modular has the API surface but no public destination data model. Wenco isn’t in the conversation. The four-vendor spectrum is wider than any procurement scoring sheet captures.
None of which is the same thing as shipping activity-based dispatch. This is not a wiring exercise between two existing systems. It is new UIs for dispatcher, planner and operator. New business flows for activity ownership, plan amendment and conformance escalation. New APIs that do not yet exist between planning and FMS. New database schemas to hold activities, sequence dependencies and spatial-SIC measurements. A different posture on what the dispatcher’s screen is for.
Whether any of the vendors are thinking this deep about the problem, or building the connector and shipping the demo, is an open question. Some will be better than others. The capability is coming whether the industry is ready or not.
What they will not sell is the operating model that makes it work. Who plans at activity granularity. Who owns the plan-to-FMS pipeline. Who keeps the plan alive through the shift. What authority Dispatch actually has, and over what equipment. How the Management Operating System changes to match. None of that is on any procurement scoring sheet. None of it can be bought.
Sites that put activity planning capability, FMS configuration, Dispatch resourcing, and Management Operating System redesign on the same project schedule will compound. Sites that buy the software and skip the operating model will end up with a more expensive version of the spreadsheet and the radio.
The five years that matter are not the ones the vendors are spending shipping the capability. They are the ones every site spends preparing the operating model that will make it useful.
Activity dispatch is the next step change in open-pit FMS
The underground sites running activity-based dispatch today are not running the same operation marginally better. They are running a different operation. The plan is alive. Dispatch is anticipating, not reacting. Conformance is real-time, not reconstructed. The supervisor with the authority to make the call is sitting in the Mine Control Centre with the underground on a series of screens, not in a pickup halfway down a decline behind a truck. That is what the capability delivers when the MOS, the technology, and the implementation all line up.
The same thing is on the table for open pit, with more fleet, more activities, and more tonnes per percentage point of conformance. The last productivity step change in open pit was Autonomous Haulage. The next one is this, and it will be bigger, because it touches every piece of equipment in the pit, not just the trucks.
The sites that get there will be the few that treat capability, MOS, technology, and implementation as one program of work, on one schedule, with one owner. They will move execution authority off the haul road and into the control room, not by removing supervision, but by giving the supervisor the only seat in the operation with the whole picture. The rest will buy the software, skip the operating model, and conclude five years from now that activity-based dispatch did not work in open pit. They will be wrong. It will have worked. They just underestimated the effort and change required to deliver the step change.