Demand and inventory planning
Retail · Food · Distribution
Forecast demand by item and location with seasonality and promotions, then set replenishment so stock-outs fall without tying up cash.
Planning and advanced analytics
Not just dashboards. Plan, predict, prescribe — and act.
Dashboards tell you what happened. VEPICT also tells you what is likely to happen, what the best response is, and can carry it out with your approval — forecasting, anomaly detection, machine learning, scenario planning and optimization, all on the same governed definitions your reports use.
- DescriptiveWhat happened?
Dashboards, paginated reports and pivots in Vista, on one semantic layer.
- DiagnosticWhy did it happen?
Drill to records, process intelligence and the knowledge graph.
- PredictiveWhat will happen?
Forecasts with intervals, anomaly flags and machine-learning models.
- PrescriptiveWhat should we do?
Scenarios and optimization that recommend a plan and price each constraint.
- AutonomousDo it — with approval.
Agents built from predictive and prescriptive blocks, run toward a goal.
Forecasting — try it
A forecast that chooses its own method, and says how sure it is
VEPICT holds back the most recent history, tries several methods on it, and picks the one that would have been most accurate — then forecasts with a prediction interval and flags the months that did not fit. This is the real calculation, running in your browser.
How each method scored on the held-back months
Optimization — try it
The best product mix inside the capacity and cash you have
Eleven product lines, each with its own contribution per tonne and its own demand. Two limits: what the works can make, and how much working capital the business will tie up. A real linear-programming solver finds the plan that earns the most — and tells you what one more unit of each limit is worth. Move the limits and watch the plan change.
VEPICT demoContribution, demand and cost per tonne by product line from the industrial manufacturer demo on demo.vepict.com (plan year; US dollars, converted at 3.75 SAR per USD).
DemandPlanned
Scenarios, solved and compared
A shadow price is the value of one more unit of a binding limit — the clearest guide to where investment pays.
Forecasting games — play
Top-down, bottom-up or middle-out: which forecast wins?
Every plan forecasts a hierarchy — a total, the groups under it, the items under those. You can forecast the total and split it down (top-down), forecast every item and add them up (bottom-up), or forecast the middle and go both ways (middle-out). Pick a plan and a situation, call the winner, then run all three on the same history and see who was closest at each level.
Top-downForecast the total; split it by each item’s share of history.
Bottom-upForecast every item on its own; add them up.
Middle-outForecast each group; add up to the total and split down to items.
Average error at each level (lower is better; next quarter, against what actually happened)
The hierarchy, next quarter
IllustrativeSimulated history (24 months) and next quarter, generated in your browser. Each forecast uses trend plus monthly seasonality fitted to its own level.
In VEPICT, Vista Predictions forecasts any measure along time from the same semantic model that holds the hierarchy — total, group or item — and Ask can run and compare those forecasts for you. This game shows how planners read the levels together.
Planning across industries
Where planning and prediction pay off
Capacity and production planning
Manufacturing · Packaging · Foundry
Plan the product mix against machines, shifts and working capital; see which limit binds and what relieving it is worth.
Workforce planning
Healthcare · Services · Transport
Forecast demand by hour and skill, roster to it, and test scenarios before overtime and burnout set in.
Budget and cash forecasting
Government · Nonprofits · Finance
Project spend against appropriations and grants, forecast cash, and see overspend months before year-end.
Asset and maintenance planning
Utilities · Energy · Mining
Predict failures from telemetry and condition, and plan maintenance and spares where they prevent the most downtime.
Revenue and renewal forecasting
SaaS · Telecom · Education
Forecast renewals, churn and enrollment from usage and engagement, and target the accounts or students at risk.
In the VEPICT platform
The analytics behind it
Vista PredictionsForecast any measure along time: automatic choice among naive, drift, linear, exponential smoothing, Holt and Holt-Winters on held-back history; prediction intervals; error against a naive baseline; anomaly flags. Download as PowerPoint, PDF or Excel.
Vista Scenarios & OptimizationLinear and mixed-integer optimization with decisions, an objective, and resource and policy constraints — typed in or generated from a dataset. Scenarios solved and compared, one recommended, shadow prices shown.
Pipelines (machine learning)Regression, classification, clustering and time series: linear and regularized regression, random forests, gradient boosting, SVM, k-NN, naïve Bayes, k-means, DBSCAN, ARIMA / SARIMA / SARIMAX, Prophet-style and Holt-Winters — seeded and reproducible.
Model registry and monitoringEvery model version traced to its data and run; monitoring after release; time-based splits so a forecast is judged on a future it has not seen.
Agent StudioAgents built from predictive and prescriptive blocks — a goal, then steps as connected cards — run in a loop until the goal is met, with human approval for any change.
Schedules and deliveryForecasts, scenario comparisons, dashboards and reports delivered as PDF, Excel, PowerPoint or CSV on a schedule, in each recipient’s time zone.
What would you plan differently if you knew a month earlier?
Bring the decision and the history behind it. We will show you the forecast, the scenarios and the plan.
Discuss planning and forecasting