What we build

Governed AI systems for energy trading and renewable operations.

Blackdown works where model outputs meet commercially consequential decisions. We combine energy-market context, deterministic engineering and explicit human control.

Origination and pricing

AI-assisted drafting and analysis for renewable origination, with economics reconciled against pricing logic the desk already trusts.

PPA and route-to-market proposalsVolume profilesCurves and shapeIndexationCertificate assumptionsException detectionAnalyst review

Certificates and portfolio operations

GoO and REGO inventory matching across registries, vintages and delivery periods, with decisions and supporting evidence prepared for review.

Inventory matchingRegistry constraintsCertificate vintagesDemand coverageTrade PnL attributionNominationsSettlement evidence

BESS optimisation and dispatch

Deterministic optimisation and governed decision support for batteries operating across wholesale, reserve and balancing markets.

Revenue stackingState-of-charge constraintsReserve commitmentsOpportunity costOverride analysisDispatch explanationForecast-versus-outturn evaluation

Governed trading workflows

Multi-step trading and operations workflows separated into model reasoning, deterministic calculation, controlled execution and named human approval.

Request triageContract interpretationPreparation of bids or nominationsApproval routingAuditabilityDecision reconstructionProduction integration

How we deliver

Clear boundaries before more autonomy.

The delivery method depends on the workflow, but the operating principles remain consistent: agree the authoritative data and calculations, define where a model can reason, preserve established authority and evaluate the result against evidence the desk already recognises.

  1. 01Scope

    Define the workflow, evidence base, principal risks and measurable outcome.

  2. 02Design

    Agree the architecture, system boundaries, control model and evaluation plan before build.

  3. 03Build

    Deliver working software against clear milestones, reviewed by the people who will operate it.

  4. 04Evaluate

    Assess performance against representative historical data and existing pricing or control logic.

  5. 05Hand over

    Deploy the system, document the controls and transfer enough knowledge for the client team to operate it.

Control comes before autonomy

Models can interpret contracts, compare alternatives and prepare decisions. Deterministic services remain authoritative for controlled calculations, while binding actions pass through the organisation's established permissions and approval process.

Start with one workflow that matters.