Define what should improve
Revenue, margin, conversion, waste, service level or another measurable business outcome.
Decision Optimisation
Compare the moves you can actually make.
Graphite Note turns model outputs into a decision problem: which feasible action is expected to create the strongest outcome under real business constraints?
The decision layer
Best model does not automatically mean best action.
A model may predict demand, churn or scrap accurately. The business still needs to decide which lever to use, where to use it and how much change is feasible.
Graphite Note structures those choices around expected impact, evidence and constraints so the output is closer to an executable recommendation than an analytical observation.
List what can actually change
Restrict the analysis to actions the organisation can execute.
Respect real limits
Capacity, budget, stock, policy, risk tolerance and other operating boundaries.
Rank feasible actions
Compare expected outcomes and make the trade-off visible.
Use optimisation when your question is
Which feasible action creates the strongest expected outcome?
Optimisation is the decision layer when there are several possible moves but budget, capacity, stock, policy or risk limits what can actually be done.
Which customers should receive the offer given a fixed budget?
Decision: allocate campaign capacity to the customers with the strongest expected incremental value within the available spend.
Where should limited stock or capacity be allocated?
Decision: distribute scarce resources across stores, regions or products using forecasts, priorities and operating constraints. See how forecasting feeds optimisation.
Which intervention should happen first?
Decision: rank controllable process changes by expected benefit, feasibility and the constraints that determine what can be executed now.
Direct answers
Optimisation in business terms.
Is optimisation the same as prediction?
No. Prediction estimates an outcome. Optimisation chooses among possible actions, often using predictions or causal estimates as inputs.
Does every use case need a complex optimiser?
No. Sometimes a ranked set of feasible interventions is enough. The method should match the decision, constraints and value at stake.
Can optimisation support assortment decisions?
Yes. Assortment is a constrained optimisation problem when the business needs to balance demand, substitution, product economics, space and strategic rules. See assortment optimisation.
Where does human approval fit?
It can remain part of the workflow. A decision object can carry the recommendation, evidence and expected impact while the final approval stays with the responsible team.
Decision IntelligenceFrom signal and prediction to a governed business action.AutoMLAutomate repetitive model-building work without losing business context.Causal Machine LearningSeparate drivers from correlations and estimate intervention effects.Predictive AnalyticsForecast demand, score customers and anticipate business outcomes.Decision OptimisationCompare feasible actions and rank them by expected business impact.Forecasting vs OptimisationSee where prediction ends and the business decision begins.Decision APIDeliver recommendations, expected impact and evidence into workflows.Embedded Decision IntelligenceAdd a decision layer to ERP, CRM, BI, workflow and vertical software.Build vs BuyA practical framework for deciding whether to build, buy or embed.
Show us the objective, levers and constraints.
We will help turn them into a decision system.