Changelog

The forecasting worksheet now shows how demand has moved over time, lets planners freeze a version of the forecast, and grades that version against what actually shipped.

A forecast is only useful if you know how far to trust it. Until now, the worksheet showed the current number and not much else. Checking last month’s forecast meant keeping a copy somewhere else. Now the platform keeps the record and does the checking.

See the trend behind the number

A new Trend tab charts the forecast against history for each product, with a confidence band around it. A product that sells the same every week and a product that swings with the season look different at a glance, before anyone reads a figure.

Freeze a forecast

Snapshots save a version of the forecast as it stood on a given day. Take one before a planning meeting, and the version everyone agreed to is still there next month, even after the live forecast has moved on.

Grade it against what shipped

The Forecast vs Actual view compares a snapshot with the quantities that really went out the door, SKU by SKU, and gives it an accuracy score. Over a few months you learn which products you forecast well and which ones need a person’s judgment.

Production requirements for every customer

The screen planners use to size new orders was rebuilt so every customer gets the same version. Order quantities and dates come pre-filled, with safety stock and projected stock-out dates alongside, so a planner starts from a proposal.

Warehouse teams can now take a load from plan to truck without leaving the platform, and every step records its own inventory movement.

When the steps of a shipment live in different places, a short count is hard to trace back to the step that caused it. Now each step is a transaction, so the history of a load is there when you need it.

Plan, pick, stage, ship

• Plan a load and confirm it from the shipping calendar
• Pick orders are created for you from the confirmed load
• Pick to staging lanes, including a load split across lanes
• Ship it, with inventory updated at every step

A two-way ERP sync

The connection to your ERP now carries production orders, sales orders, products, and inventory policies in both directions, and routes each change as it happens. Nothing has to be re-keyed to keep the ERP and the floor in agreement.

A scheduling agent that asks first

A new agent watches the production schedule and proposes reschedules when something changes. It never moves the schedule on its own: every proposal goes to a person for approval. When a quality hold is approved, the agent can propose a reschedule for the affected line straight away, so the hold and the plan stop disagreeing.

Paper forms, imported

Upload a PDF, a photo, or a Word document of a paper form you already use, and the platform turns it into a draft digital quality form. A person reviews the draft before it is saved, so nothing reaches the floor unchecked.

Planners can now sequence production orders on a board, dragging them between areas, lines, and assets. The rules for each product line are checked on the drop.

The board puts the planner’s sequencing on the screen. Move an order to a line that cannot run it, and the board tells you before the order lands there.

Run Schedule board

• Drag orders by area, by line, or by asset
• Product-line rules are enforced as you drop

Suppliers as a managed record

Suppliers used to be a name typed on each purchase order, so the same supplier could be entered more than one way and its history was hard to pull together. They are now a single managed record that every purchase order points to. Existing purchase orders were moved over automatically.

Inventory on physical bins

Holds, transfers, and moves now work on the physical bin a quantity sits in. This continues the inventory work from the last release, and it is what makes location-level counts line up with the shelf.

Inventory can now exist across several physical bins at once instead of being tied to one location. This is the largest change in the release, and the foundation for counts you can trust at the shelf.

A lot rarely sits in one place. Part of it is on the line, part in the warehouse, part in a staging lane. When the system only knew the lot, a picker had to know the rest. Now the platform knows where each quantity is.

Every movement is a transaction

Consumption, adjustments, receiving, and picks all run through an audited transaction on the bin record. If a count is off, the trail shows which movement made it so, and who made it.

Location rules for everything you stock

Rules about where things may be stored now apply to raw materials, consumables, and packaging as well as finished goods.

The shipping calendar is the warehouse home screen

The shipping calendar is now the first screen the warehouse sees. Orders that do not have a carrier yet get their own exception path.

The first AI agent on the platform is live. It watches reject rates on the floor, and when something looks wrong it drafts a likely cause and sends a hold proposal to a person. Nothing is held until someone approves it.

A rising reject rate is easy to see in the next morning’s report, after a shift of scrap. The agent looks every fifteen minutes, so a bad run is caught inside the shift.

How it works

• It checks the assets you configure every fifteen minutes for reject rates that are statistically out of pattern
• It drafts a root-cause hypothesis from what it can see
• It sends a hold proposal to the Approval Queue for a person to accept or decline

Shipping calendar

A new weekly calendar shows carriers, loads, and inbound orders on one grid. Drag a load to a different day and the plan updates.

One panel for the operator

Operator screens now bring the job queue, start and stop controls, and the quality forms a job needs into one panel, so an operator is not switching screens mid-run.

Operator screens now know which machine they are on. The order picker filters to the orders for that machine or line, and picks the order that is already running.

On a floor with several lines, a shared order list is mostly other people’s work. Filtering to the machine takes the wrong pick off the table.

Knowledge search, as it happens

The built-in knowledge search now streams its answer as it is written. Two more changes make answers easier to check:

• Ask to see the query trace, and it shows how the answer was found
• Ask a question as of a past date, and it answers from what was true then

Lot, order, and location names on screen

Several screens now show the lot, order, and location names people use instead of internal ID numbers.

Knowledge search was rebuilt on a new retrieval engine. Every answer now shows where it came from and how confident the platform is in it.

On a floor where the answer decides whether a lot ships, people need to see the source before they act on it. Now they can.

Sources and confidence on every answer

• Each answer cites the documents and records it drew on
• A confidence badge shows how strong the match was

The trace workspace comes first

The trace and genealogy workspace is now the first screen for tracking a lot’s history. It shows the lot’s genealogy, lets you put it on hold from the same screen, and uses names people recognise throughout.

Lots assigned as the job runs

A new operator widget assigns material lots to a job in real time and tracks consumption automatically, so the genealogy is written while the work happens.

Planners can now build a what-if scheduling scenario, save it, and see how it plays out before touching the real schedule.

A scenario runs against the same capacity math as the live schedule and comes back with a verdict, so a planner can rule out a bad idea before it goes on the real schedule.

Try it before you commit

• Model demand as a single date, a spread, or a recurring pattern
• See a Feasible, Tight, or Infeasible verdict across the horizon you set
• Get the likely bottleneck, and compare two scenarios side by side

One widget for putaway

A new operator widget puts job-order selection and putaway execution in the same panel. An operator picks the job order, works the putaway queue with simple skid counts, and the job order’s actuals update as they go, no switching screens to reconcile the two.

Investigations get their own module

Root-cause investigation and corrective-action work used to live as a tab inside Quality. It’s now a module of its own, with its own dashboard, case list, issue list, and kanban board, so a case doesn’t get lost inside someone else’s queue.

Alongside that, a new quality gate can require a form to be submitted and pass before the next production step is allowed to start.

The platform’s navigation was rebuilt from the top down: a persistent top bar, a command palette, and a slim module rail in place of the old sidebar and stacked tabs.

Routes and permissions did not change. What changed is how much of the screen they take up to get to, and how fast you can jump somewhere without hunting through menus.

What’s in the new top bar

• A command palette to jump straight to a screen or action by typing
• One switcher for enterprise and site, instead of two separate controls
• Notifications and your account menu in a fixed, predictable place

Page header height dropped from around 110 pixels to around 54, so more of the screen is the work itself.

Sales forecasts now refresh themselves

A background job now refreshes every active product’s sales forecast automatically, twice a day. Any manual lock or override you’ve set on a forecast is preserved.

A dedicated workspace for tracing a lot’s history shipped, alongside a new module for logging issues, causes, and countermeasures through to close.

Tracing a lot and working an investigation each get a home built for the job: one workspace for following a lot’s history, one module for working an investigation from open issue to closed case.

Forecasting down to the customer

Sales forecasting can now be viewed at the customer and SKU level, with roll-up to the aggregate forecast you already had. A planner can see exactly which customer or product is driving a change in the number.

Quality holds become their own workflow

Putting a lot on hold is now a dedicated workflow rather than a side effect of another action, which makes it easier to see what’s on hold and why.

Materials and purchasing

• Materials now reorder based on a configured reorder point instead of a fixed schedule
• Purchase requisitions can be converted directly into purchase orders

Every dashboard metric now explains itself. Hover over one and you see what question it answers and how it’s calculated, so the number in the meeting is explained.

A new utilization widget joins the dashboard too, with a headline number, a trend line, and a breakdown by asset, and every major widget now has its own time-period selector instead of one setting for the whole page.

A calendar view for the schedule

The production schedule now has a calendar view: one card per shift, listing the products scheduled on it. It sits alongside the existing view.

Tell the scheduler what to run first

A new product-level sequencing priority lets a planner rank specific products to run earlier, independent of the other planning rules already in place.

A summary at the top of recommendations

Production recommendations now open with a summary of the highest-moving products and a color-coded health status for each, so a planner sees what needs a decision before scrolling into the full list.

A new shipping planning board shows a planner the sequence of open shipment requirements for the product currently being produced, built automatically the moment a sales order is confirmed.

The platform builds the shipment queue from confirmed orders: earliest ship date first, then oldest order, with the applicable pallet rule already matched.

A production order can now carry its own assigned output lot, with every posting after that routed to it automatically.

A planner assigns the lot up front, and history stays put: changing the assignment only affects postings still to come, never what’s already recorded.

Fleet and equipment tracking

A separate module for managing fleet and equipment shipped alongside it: a dashboard of reservations, a schedule for pickup and return, a map of where equipment actually sits, and a list of assets with maintenance status.

Materials and forms

• Materials configured with time-based consumption now appear consistently across every materials-planning screen
• The quality form builder’s conditional-visibility rules now support multiple conditions combined with AND logic

A sales order can now run all the way from allocation through picking, load building, and bill-of-lading generation in one configurable pipeline.

Because the pipeline is configured per site, a warehouse can define its own sequence of steps. A readiness panel on the sales order screen shows every step and its live status, so a planner can see exactly where an order sits without asking the warehouse.

Quality gets granular permissions

Quality management moved from one blanket permission to permissions set per screen: dashboard, records, schedule, holds, and root-cause each control access separately now.

A dashboard you can rearrange

The executive dashboard was rebuilt on a widget system. An admin can enter an edit mode, add widgets from a catalog, and drag them into the order that matches how the team actually works.

A new forecasting engine now generates a sales forecast and a 6-month inventory projection, and a planner can edit or override either one when their own judgment says otherwise.

What the projection shows

The inventory projection looks 6 months out and flags projected stockouts and safety-stock coverage along the way.

Forms that adapt to the answer

The quality form builder gained field-dependency logic, default values for text and dropdown fields, and a note field type for instructions that only need to appear in context.

Job orders that close themselves

A job order can now complete automatically at shift end once production has occurred, instead of waiting on someone to close it by hand.

Demand planning shipped as its own capability, giving planners a forecast-driven starting point for production planning instead of building it from scratch each cycle.

Quality management was rebuilt the same release, on a new architecture connected directly to live production data, and its earlier version was retired once the rebuild was live.

What came with the quality rebuild

• Statistical process control gained dedicated chart records for tracking measurements over time
• Custom versus standard user roles are now visibly badged on user screens

Suggestions from continuous improvement

The continuous-improvement module’s suggestions were improved for relevance.

Also in this release

Purchase order management shipped, alongside a simplified production-schedule screen and a command center view for production planning.

The platform’s built-in knowledge assistant shipped for the first time: upload the documents your team already works from, and ask it a question in plain language.

Process knowledge now lives in a collection the assistant can search, with a quick-access widget so a question doesn’t mean leaving the screen you’re on.

What shipped in this first version

• Document upload and collection management, with an admin screen to organize them
• A search interface with suggested questions to get started
• A quick-access widget available from wherever you’re working

A first materials-planning module

A first version of materials planning also shipped, calculating material requirements directly from demand, without a full bill of materials yet.

Also in this release

A light-mode theme joined the existing dark theme, and OEE metric labels and time formatting were cleaned up for clarity on the metrics screen.

The automated scheduler now logs its own processing time and how many orders it scheduled on every run, so its efficiency can be tracked over time instead of judged by feel.

Login and password reset both gained email verification.

Also in this release

• OEE and utilization metrics now use each asset’s configured targets and exclusion flags consistently across every screen that shows them
• Dedicated accounts were set up for connected integrations, separating that traffic from any individual user’s login

The order detail screen was redesigned with tabbed navigation between details and batches, so reviewing an order and its batch breakdown no longer means two different screens.

The production schedule gained a matching batch view, toggle-able alongside the existing order view, with the same filters carrying over between the two so switching views doesn’t mean resetting your search.

Fewer manual steps behind the scenes

Production steps are now created and kept in sync automatically as orders come in from a connected system, instead of needing a manual pass to line them up.

Send a support ticket from the platform

You can now send a support ticket directly from the platform, and search was fixed and improved across product recipes, demand, and product master screens.

The platform adopted a new typeface across every screen, chosen for a more consistent and accessibility-compliant reading experience.

Quality management moved out of a separate prototype environment and into the main application for the first time this release, so it now lives alongside the rest of your operation instead of off to the side.

Also in this release

• Long selector lists on the OEE metrics screen replaced with dropdowns for areas, lines, and assets
• Asset-state history screens gained new filtering and grouping options
• Temporary passwords now expire after 7 days

The scheduler can now account for changeover time, the gap it takes to switch a line from one product to the next, instead of treating every changeover as instant.

Changeover times are defined once, in a matrix of product to product, and can be imported in bulk. The automated scheduler reads that matrix and factors real changeover time into the schedule it builds.

Tell it what to prioritize

Two new screens let a planner set explicit rules: a planning-sequence screen for how production orders should be ordered, and a planning-priorities screen for the priority classifications that sequencing draws on.

Capacity that reflects reality

Assets gained a parallel-job-capacity setting, and shifts gained a percent-of-capacity setting, both now factored into automated scheduling, so a schedule better reflects what a line or a shift can actually carry.

A dedicated changeover screen shipped this release, the first place to track and manage the time it takes to switch a line between products.

It’s a small screen with a specific job: set changeover times once, and let the rest of production planning read from it, rather than a planner keeping that knowledge in their head.

Also in this release

• Admin configuration tabs are now restricted to the right permission level for the first time
• Order and job-order numbers were cleaned up across the interface for readability
• Executive reporting gained tooltips on late orders, linked directly to their order detail screens