Planning the ingredients plant around what is expensive to stop: the reactor, the dryer and the extruder.
In an ingredients plant, the assets that drive cost are the continuous ones: the reactor, the spray dryer, the compounding extruder. Every changeover costs — time reaching temperature and pressure, off-specification product during start-up and, sometimes, material that has to be scrapped. The iLEAN Planning Agent builds the month's plan with long campaigns and the fewest possible changeovers — starting from the spreadsheets you already have.
The plan is built looking at demand — and the reactor and the dryer pay for every changeover.
The reactor, the dryer and the extruder of an ingredients plant do not switch products for free. Every changeover comes with an invoice the spreadsheet never sees:
- The time to reach operating conditions. Getting to working temperature, pressure or humidity takes time — with the equipment running and no conforming product coming out.
- The transition product. Until the process stabilizes, the product comes out off specification: moisture, particle size or composition out of range. Depending on the ingredient, that means rework or scrap.
- The changeovers that make a mess. Switching from one product to another can require long purges or cleans; a fragmented plan multiplies those changeovers and leaves the equipment in changeover instead of in production.
And who decides how many changeovers happen in a month? The production plan. With a spreadsheet, the planner barely gets to a feasible plan, let alone to comparing "do I maximize throughput or minimize expensive changeovers?". And redoing it when demand changes takes days.
Connect absorbs your spreadsheets; the Agent builds the plan that respects the reactor and the dryer.
The iLEAN Planning Agent was born planning assets that genuinely cannot stop — continuous assets in heavy industry, where continuity is a hard constraint. That same machinery is dialled down to what your plant needs: strict continuity if you decide so, or a heavy penalty on every start-up so the engine minimizes them without forbidding them.
The user drops in their spreadsheets. The agent builds the model of the ingredients plant — reactor, dryer, extruder, items, changeovers — and returns an editable receipt. The engine solves with the constraints of the continuous assets. The human compares three scenarios and approves with a stamp.
The specific iLEAN piece for your ingredients plant:
- Connect — absorbs the chaotic documents the plant already has: the production spreadsheet with its thousand tabs, the commercial demand that arrives by email, the rates per SKU and per line. No mandatory template and no integration project up front.
- Planning Agent — builds the planner's database on its own (assets, lines, items, constraints, demand), asks only when something blocking cannot be derived from the documents, and solves with a constraint-based optimization engine. For the reactor, the dryer and the extruder it applies what matters: minimum campaign length, sequencing that groups compatible products to minimize purges and expensive changeovers, continuity or a start-up penalty depending on how it is configured, and the affinity of each product to its line. And it plans the full cascade — reaction → drying/extrusion → milling → packing — flagging the limiting input whenever it bites.
- Three scenarios, not one plan — maximum output, tight inventory and stability (long campaigns, minimum changes). Side by side, with KPIs and the differences explained one by one.
The discrete assets — batch milling, packing, bagging — are planned by compressing the same production into fewer hours and placing them, if your contract allows it, in the cheap energy bands, without leaving the continuous assets without an outlet.
The approved plan is exported to the spreadsheet in the format of the production program the plant already uses. The team's routine does not change on day one — what changes is how long it takes to get to a good plan.
Plan in Excel vs. plan with the Planning Agent
| Aspect | The planner's spreadsheet | With the Planning Agent |
|---|---|---|
| Time to build the plan | Days | One morning, with the agent building the model |
| Reactor / dryer changeovers | Whatever comes out of fitting it by hand | Minimized by grouped sequencing |
| Off-spec transition product | Whatever falls out at each changeover | Fewer changeovers, less off-spec product |
| Purges and cleans between products | Multiplied by short campaigns | Grouping that minimizes dirty changes |
| Milling and packing | Inherited shifts running part-loaded | Compressed and, where it pays, moved to off-peak hours |
| Strategic alternatives | A single feasible plan | 3 scenarios compared with KPIs + what-ifs |
| Replanning after a surprise | Another few days | Minutes: recalculated and compared |
| Format of the approved plan | The usual spreadsheet | The same usual spreadsheet — just exported |
Impact estimate for your plant — to be validated with your numbers.
The block below is an estimate to be validated with the specific data of your plant. We put it forward so the committee has an order of magnitude; we refine it during the diagnostic.
- Ingredients plant with a reactor, a spray dryer or a continuous extruder plus discrete assets (milling, packing), monthly plan built today in a spreadsheet by one person.
- First planning session with the agent on the existing documents — with no prior data project. First comparable plan in one morning.
- Reduction in expensive changeovers and off-specification transition product expected at ≥30% versus the manual plan, thanks to grouped sequencing the engine finds and the spreadsheet does not. (Conservative range — estimate to be validated.)
- Indicative payback between 4 and 9 months, depending on the consumption of your continuous assets, the cost of the transition product and the planning hours freed up.
The defensible technical anchor comes from the planner's best-documented real case: a high-demand industrial plant in heavy industry — with a strictly continuous asset — whose agent built the complete model on its own (one kiln with 3 lines, 128 items with their production rates, 5 operational constraints and around 4,600 tonnes of demand) from a 114 MB zip file containing an 18-sheet spreadsheet, and delivered 3 optimized scenarios in under an hour. The same class of constraint engine, dialled down to what a reactor or a spray dryer needs.[1]
And the operations director's reasonable doubt
"What if the plan the AI proposes cannot actually be run in my plant?" — that is why there is human verification at every gate. The intake receipt shows everything the agent understood from your documents and is corrected inline before anything is solved; the scenarios are compared with the differences explained; and the plan only goes to production when the person in charge approves and stamps it. If the demand does not fit in full, the system does not say "infeasible": it delivers a partial plan with its coverage and points out exactly what has been left out, so that a person decides. The AI proposes; the plant signs.
[1] Real iLEAN case, high-demand plant in heavy industry — verified pilot figures: 114 MB of documents, 18 sheets, 128 items, ~4,600 t, 3 scenarios in <1 h, 99.5% coverage.
What people ask about planning production with an agent
Are the reactor and the dryer "continuous" assets like a tile plant kiln?
They are similar because of what stopping costs, not because they break. A tile kiln cracks if you stop it; a reactor or a spray dryer will not break, but every changeover is expensive: time spent reaching temperature and pressure, off-specification product during start-up and, sometimes, material to be scrapped. The planner lets you choose how to treat them: strict continuity, or a heavy penalty for every changeover — and the plan comes out with the long campaigns these assets appreciate.
What does the agent need to build the plan for my ingredients plant?
The documents you already have: the production spreadsheet, the demand, the rates and changeover times per product, the sequence compatibilities and the quality constraints. No mandatory template. The agent builds the model and returns an editable receipt with every decision it took written down, before it solves anything.
How does it minimize expensive changeovers and purges?
By sequencing compatible products into long campaigns, so that products sharing the same conditions are grouped together and the changes that require a purge or a clean are kept to a minimum. It is a data-configured constraint — minimum campaign length and a penalty per changeover — that the optimization engine respects while looking for the plan that reduces changes without missing demand.
And milling and packing, which I can compress?
Those are discrete assets: their cost is proportional to the hours they run. The engine compresses their production into fewer hours and, if your electricity contract has time bands, places them in the off-peak windows, using intermediate product buffers so the continuous assets never run out of an outlet. The scenarios compare the compressed plan against the conservative one.
Who approves the plan: the AI or the plant?
The plant, always. The AI proposes: the intake receipt is editable, the scenarios are compared side by side, and the plan only goes to production when the person in charge approves and stamps it. The approved plan is exported to the spreadsheet in the format the plant already uses.
Tell us your case and in 48h we'll send you the estimated ROI of this AI project for your plant.
We work on your plant's real data, not ours. Diagnostic with no commitment.
Request estimated ROI in 48h ‹ All ingredient cases See food industry