From scattered processes to one operating system.

At Rekeep's scale the recipe is the basis of food cost, clinical nutrition, material requirements, purchasing and production planning. Rekeep needed one standardized model for designing recipes and technology that every downstream process could rely on.

Problem

The previous approach did not reproduce the full production process, least of all the losses and yield changes at cleaning, processing, cutting, cooking, chilling and portioning. A kilogram of raw material is not a kilogram of product, so requirements, food cost, warehouse issues and the plan-versus-actual comparison all came out wrong.

Analysis and mapping

We mapped the real production process, not just what a dish consists of, but how it actually comes about: raw material, operation, change in weight and yield, semi-finished item, next step, finished good. We tied the recipe to production schemes and work centers.

What we built for Rekeep

• Advanced Recipe Engine with nested recipes and semi-finished items
• Yield & Loss Engine covering every stage of the process
• Production Schemes tying a recipe to its sequence of operations and work centers
• Automatic recalculation of nutrition values and allergens
• Recipe versioning with impact analysis: where it is used and what a change affects
• Alternative ingredients compared on nutrition, allergens, diet fit, technology and cost
• An AI Recipe Cost Optimization layer suggesting cheaper ingredients within dietary limits

Rekeep connects information from a large number of hospitals with the capacity of its kitchens and central plants. The first question is how many meals will have to be produced, and only then what, where, when and in what order, across several hundred sites and multiple plants at once.

Problem

Running the process only off firm orders pushes decisions too late. On top of that come swings in patient numbers, diet mix, differences between wards, days of the week, seasonality, holidays and order corrections, and aggregating data from several hundred sites by hand caps how far ahead purchasing and production can be planned.

Analysis and mapping

We mapped the process from the hospital up to the production plant and built a planning module that reproduces the flow: forecast demand, aggregate demand, calculate material requirements, build the production plan, schedule production, monitor execution, replan.

What we built for Rekeep

• Hospital Demand Forecast: site-level forecasting before the order cut-off, from patient history and time dependencies
• Production Demand Forecast: demand aggregated across diets, meals, dishes, recipes, semi-finished items and raw materials
• Fusion- and Transformer-class models compared so the best prediction wins for each data series
• Production Planning Engine: plan by work center, production and packing schedule, resource load, simulations and replanning

The production plan raises the next question: how much of which ingredients to buy, from whom, at what price and by when. At Rekeep's scale purchasing is one of the biggest sources of both savings and uncontrolled cost. It should be an automatic consequence of what the organization intends to produce, not another place to type orders in by hand.

Problem

No single source of current price lists, supplier knowledge scattered across teams, a large volume of manual operations, no link between the production plan and purchasing, limited price and product comparison, and no unified view of the history with a supplier.

Analysis and mapping

We mapped the full flow: production plan, material requirement, inventory coverage, purchase requirement, supplier selection, purchase order, delivery, receiving, supplier evaluation.

What we built for Rekeep

• Automatic Purchase Requirements: the plan resolved through recipes and current inventory into a ready buying list
• Supplier 360: a central supplier profile with products, terms, price lists, lead times, minimums, quality and claim history
• Dynamic Price Lists: fast price list import and updates with validation and supplier-index mapping
• Supplier Integrations: digital data exchange with the largest suppliers, a simplified portal for the smaller ones

A correct order still does not guarantee cost control. The critical moment is when the goods physically arrive at the plant and someone has to verify that what was ordered actually turned up, in the right quantity, at the right price, quality and time. Receiving is where control over ingredient cost and supplier quality is either established or lost.

Problem

The analysis showed incomplete delivery records and no uniform digital receiving process: a delivery did not always reach the system immediately, there was no full receiving history or consistent quality assessment, and discrepancies were sometimes logged late.

Analysis and mapping

We first straightened out the target process: planned delivery, arrival, document capture, PO matching, quantity check, quality check, lot and expiry registration, accept, partially accept or reject, inventory update, supplier performance data. Only then did we build the software.

What we built for Rekeep

• Delivery Schedule: every delivery due on a given day in one view
• Digital Receiving on a tablet: full, partial, rejected, split-lot and claimed receipts
• OCR Document Capture matching document data against the order
• PO vs Delivery Control: shortages, overages, unordered goods, price differences, wrong product, expiry and temperature issues
• Supplier Quality Data building a record of compliance, quality and on-time performance

Correct inventory still does not solve the flow of materials to production. What counts is not only what sits in the warehouse but what should be issued, when, to which station, to which work center and against which recipe, and with many production runs going on at once, managing that by hand means delays, mistakes and lost control over real consumption.

Problem

Without a digital issue process it is hard to say which material should go out next, in what quantity, to which work center and station, for which recipe, when and by whom. Material issued too early takes up space on the floor, issued too late it stops production, and issues logged after the fact strip inventory of its credibility.

Analysis and mapping

We connected warehouse logic straight to production planning: production forecast, production plan, recipe schedule, material requirement, issue schedule. An issue stops being an independent warehouse task and becomes part of the production plan.

What we built for Rekeep

• Intelligent Issue Queue: an ordered issue timeline with time, material, quantity, location, work center, station, recipe and order
• Mobile Warehouse Tablets: issues completed on a handheld, from opening to pick confirmation and close
• Real-Time Inventory Update feeding production, planning, purchasing and controlling instantly
• Digital Internal Documents giving every movement a full trail
• Additional Material Requests turning a shortage call from the floor into a warehouse task rather than a phone call

Once the plan is set and the materials are delivered, physical production begins. Rekeep needed a system that walks operators through the plan and gives the manager full visibility of the floor: one environment connecting plan, work center, operator, recipe, material, time, yield, loss and the finished result.

Problem

Without a digital MES a production manager cannot tell which recipe has started, at which work center, how far it has got, whether the operator has every ingredient, how long an operation takes, how much product was actually made or how much material was lost. Operators, meanwhile, need current instructions and tools that spare them recalculating recipes by hand.

Analysis and mapping

We mapped the real life of a recipe on the floor: materials available, operation start, instructions followed, time recorded, weighing, step closed, handover to the next work center, final yield. Every recipe was tied to the plan, work centers, stations, operators, materials, operation times and expected yield.

What we built for Rekeep

• Live Production Control: the status of every recipe at every work center in real time, with progress and delay risk
• Digital Work Instructions on tablets with recipe, materials, quantities and operation order
• Material Readiness Check before a recipe starts
• Start/Stop Tracking recording actual time, operator, station and work center
• Smart Production Calculator that scales a recipe in both directions
• Additional Material Request raised from the tablet
• Weight & Yield Registration comparing actual against expected weight at every step
• Full Production Traceability: who, what, where, when, how much, from which material and with what result

Executing production is not the last step in the process. It is at the same time the source of new data for the next plans and forecasts. The system was built on the premise that every production day yields data that can improve the next one.

Problem

A forecast and a schedule are only as good as the data behind them. A system that knows only theoretical values cannot say how long a recipe really takes, what the actual yield is, where the biggest losses arise, which work centers run faster or how often a replan is needed.

Analysis and mapping

We treated every production operation as a data point. The system compiles forecast against actual: planned and real start, finish, operation time, yield, consumption and volume.

What we built for Rekeep

• Forecast vs Actual, compiled automatically once production closes
• Historical Production Data at recipe, work center, station, shift, operator, day and plant level
• Prediction Feedback Loop returning actuals to planning and prediction, closing the loop
• Process Intelligence spotting repeat delays, underestimated times, yield variance, problem recipes and overloaded work centers

In hospital foodservice packing is a critical stage: sequence, portion weights, pack type, the layout of the components, consistency with diet and recipe and correct labeling all have to hold. With that many diets, dishes and variants, managing the packing order by hand raises the risk of errors and stoppages.

Problem

Without a digital module it is hard to hold the optimal packing sequence, stay aligned with the plan, repeat portion weights, place components correctly, give operators unambiguous instructions, control operation time and trace the finished product, all at once.

Analysis and mapping

We connected packing straight to the production plan and the real execution status of recipes: what is ready, what should be packed, in what order, into which pack, at what portion weight and under which label.

What we built for Rekeep

• Intelligent Packing Schedule prompting what to pack and in what order, updating as production progresses
• Start/Stop Packing recording actual operation time
• Digital Packing Instructions with portion weight, pack type, tray compartment and layout order
• Visual Reference: a photo of the correctly packed product
• Smart Label Management for single packs, trays, bags and cases tied to a lot
• Packing Traceability recording recipe, lot, operator, station, times and pack type

The greatest value of digitization appears when data from recipes, the warehouse, production and packing is analyzed together. No module is a separate application: each one feeds a shared data layer. Rekeep needed an answer to one question: what happened to the cost, the ingredient and the recipe, from plan all the way to finished product.

Problem

In a scattered environment data answers only local questions: the warehouse knows what it issued, production knows what it made, packing knows what it packed and purchasing knows what it ordered. The picture of the whole process is missing.

Analysis and mapping

We connected data across the chain (recipe, forecast, purchasing, receiving, warehouse, issue, production, packing) so the process can be analyzed end to end.

What we built for Rekeep

• Plan vs Actual across every step of the process
• Yield Analytics at recipe, step, work center and plant level
• Production Time Analytics
• Material Consumption Analytics
• Packing Performance
• Supplier quality data tied to how material was actually used
• Management Dashboards for managers and the board: production, food cost, variances, waste, consumption, forecast accuracy and process performance