Evaluating Unified Surface Mine Planning Tools

Open-pit planning is three jobs that pretend they are one. The scheduler says when a bench gets dug. The blend plan says which ore can hit the crusher or the plant without blowing a spec. The haulage model says whether the trucks and the dumps can actually move that material in the hours you have.

On a lot of sites those three jobs still sit in different files, sometimes in different products. Each file can look clean. The week still comes apart on Tuesday afternoon when the crusher is hungry, the dump is full, and the only ore that meets spec is two lifts away.

In early September 2026, Deswik, part of Sandvik, launched Deswik NOVA as a surface mine planning product that, according to the company, brings mining, blending, and haulage into one environment. That launch is the reason this piece exists. It is not a reason to buy anything. It is a reason to write down what "unified" has to mean on your pit before anyone changes the workflow.

Unified does not mean one login. It means a change in one constraint shows up in the other two without a hand export.

If you tighten a blend spec, the schedule should move, and the truck hours should move with it. If you lose a dump, the blend should tell you which parcels are now stranded. If the plant is down for a shift, you should see the stockpile and the dig sequence change in the same sitting, not after someone rebuilds a spreadsheet overnight.

That is a higher bar than "the modules are in the same suite." Suites have been bundled for years. Planning engineers still spend the evening reconciling a schedule file against a blend file against a haulage file because the objects are not the same objects. A bench in the scheduler is not automatically the same parcel in the blend, and neither is automatically the same load in the haulage network.

Ask the vendor, any vendor, to show that identity. Same block, same period, three views. If they have to export, rename, or re-aggregate to get there, you do not have a unified plan. You have a faster copy-paste.

The other meaning of unified is ownership. Someone on site has to be able to re-run the plan next month without the person who built the first model. If the only person who can connect schedule, blend, and haulage is a visiting consultant, you have rented a picture of integration. You have not bought the capability.

Sequential planning is not incompetence. It is how most pits grew up. Optimise a pit shell. Design pushbacks. Schedule the dig. Then ask blending what the plant can take. Then ask haulage whether the trucks can keep up. Each step uses the last step's output as if it were still true.

The break is timing. Blend and haulage constraints often only become visible once the sequence exists. By then the planning window is already compressed. Operations patch the day. The long-term plan quietly drifts.

Amanda Forbes, a principal mining consultant at Deswik, put the same failure in vendor language in a July 2026 Australian Mining feature. According to Forbes, when mining, blending, and haulage are handled in separate systems, each piece can look valid on its own without working as a whole, and teams burn time moving data instead of improving the plan. Treat that as a description of a common site problem, not as proof that one product fixes it.

You can test the problem on your own last quarter without buying anything. Pull last month's planned tonnes, the actual plant feed grades, and the actual truck hours from fleet management and dispatch. If planned ore, delivered spec, and truck hours cannot be laid on the same calendar without an argument, sequential planning is already costing you. A new tool that cannot ingest those three histories is not going to close the gap.

The other break is people. Forbes also said, according to Deswik, that planning teams are being asked for more scenarios with fewer experienced people. That pressure is real on a lot of Australian and Canadian pits. It does not follow that a guided workflow is the right answer. Sometimes the right answer is fewer scenarios, a clearer owner, and a model that a graduate can re-run on Friday.

Use this before a trial, not after the licence draft arrives.

1. Same objects, three views. Pick one bench from last month. Ask to see it as a schedule task, a blend parcel, and a haulage source in one session. Time the clicks. If the demo uses a toy model, insist on your block model and your dump network for the proof.

2. Change one lever, watch the other two. Cut plant throughput by 15 percent for a week. The schedule and the haulage should move without a rebuild. If the vendor has to "come back to you," the integration is a slide, not a workflow.

3. Horizons. Long-term, budget, and weekly plans have different jobs. Ask whether they share a model or only a brand. A tool that shines at life-of-mine and falls over on a seven-day look-ahead will not survive contact with the shift boss.

4. Export and audit trail. You will have to defend tonnes, grades, and strip in front of a manager who was not in the room. Can you export the assumptions, the constraints, and the result in a form a competent person can re-run? A picture of a Gantt chart is not an audit trail.

5. Haulage versus dispatch. Planning haulage is a forecast of truck hours and destinations. Live dispatch is what fleet systems actually ran. Ask how the planning tool's haulage result is compared to last month's actuals. If that loop does not exist, you will plan a fleet you do not have.

6. Blend ownership. Who sets the spec: geology, processing, or the planner? A unified tool that hides the spec inside a default will look clever in the demo and fight the metallurgists in week two.

7. Who runs it in six months. Name the site owner. Name the backup. If both names are at head office, the pit will go back to spreadsheets the first time the model will not open.

8. What you will stop using. Integration that adds a third scheduler above the two you already have is not integration. Write down which file dies if this works.

None of those questions pick a winner. They stop you buying a story.

Deswik NOVA was announced around 1-2 September 2026 during Sandvik's Future of Mining event in Tampere, Finland. Independent coverage ran in International Mining and Canadian Mining Journal. Deswik had already described the same idea in a July 2026 Australian Mining feature.

According to Deswik, NOVA is a surface mine planning solution that puts mining, blending, and haulage in one integrated environment, with a guided workflow from pit to product. Glenn Wylde, Deswik's chief technology officer, said surface mine planning has become more interconnected and more data-intensive, and that NOVA is meant to give planners more visibility. Riku Pulli, Sandvik's president of digital mining technologies, said customers need planning tools that are more connected and easier to use. Deswik also said the product is designed to support timely ore delivery and better use of haulage capacity.

Those are vendor claims. They are useful as a statement of intent. They are not evidence that your pit will get more scenarios, cleaner blends, or fewer truck hours.

MSR already lists an independent profile for Deswik.NOVA under mine planning and design. The directory copy describes it as a unified open-pit planning tool combining scheduling, blending, and haulage optimisation. That is a product description, not a ranking and not a recommendation. There are no community reviews on the profile yet. Until someone who has run it on a real fleet writes that down, treat the launch as a signal that the category is moving, not as proof the workflow problem is solved.

Mount Gibson Resources is named on Deswik's own site as a site that adopted NOVA inside a broader Deswik CAD and optimisation setup. That is a vendor-published example. It tells you the product is in the field. It does not tell you whether the three-way identity of schedule, blend, and haulage held up after the first quarter.

If you evaluate NOVA, or any other unified surface planner, run the checklist above on your data. Ask for the same-object demo. Ask for the 15 percent plant-down scenario. Ask who on your payroll can re-run it.

A surface planning tool does not replace geological modelling, and it does not replace the dispatch board. It sits between a block model you trust and a shift you can actually run.

On the way in, you still need a governed model: tonnes, grades, densities, and a QA/QC trail that will survive a JORC or NI 43-101 question. If that model is still moving every fortnight, you are buying a machine you will re-run constantly. That can be the right buy. It is a different buy from a one-off study tool.

On the way out, operations will still break the plan. Weather, a shovel down, a dump closed. The useful test is how fast the planning model can absorb that break without becoming fiction. A unified environment that takes a day to refresh is not more unified than three files that a planner can update before pre-start.

Deswik.Planning, Deswik.CAD, and the rest of that suite still exist as separate products on many sites. A launch that "connects planning more closely with the broader mining workflow," as Deswik put it, only matters if the site can point to the file that is now redundant. If the answer is "we will keep all of them," you have added a layer.

The directory is there to compare what tools claim to do, not to tell you which one to sign. For the wider field, start at the mine planning and design category and keep the questions in this article next to the demo, not in the brochure.

The pit does not care that the vendor had a launch in Finland. It cares whether Tuesday's trucks, Wednesday's blend, and the month's strip still add up when you change one number.