Objectives, priorities, constraints and success criteria are captured as part of the decision context.
Confidential development and validation environment
Decision Intelligence for Complex Engineered Products
Strategyfinder is developing a product independent Decision Intelligence platform that connects customer intentions, engineering geometry, product knowledge, evidence, uncertainty and accountable human review within one structured digital environment.
The Tensevo Configurator is the first industrial showcase used to validate and further develop the platform under real project conditions.
The Platform
From product configuration to accountable decisions
Conventional configurators describe what can be selected or priced. Strategyfinder adds a formal decision layer that shows whether a proposed solution is sufficiently supported, where uncertainty remains, what is blocking progress, which information is missing and who must approve the next step.
Geometry, location, technical values, product rules, costs, manufacturing logic and project status are connected.
The platform distinguishes known facts, assumptions, missing evidence, critical blockers and unresolved decisions.
Technical, commercial and legal decisions remain assigned to named human roles and documented review states.
The platform does not merely generate a variant. It explains why a variant may be suitable, which evidence supports it, what remains uncertain and what must happen before the project can responsibly proceed.
How It Works
One structured path from intent to decision readiness
Each project combines product neutral geometry, product specific logic and a formal model of evidence, uncertainty and responsibility.
Capture
Collect customer goals, site conditions, geometry, constraints and available technical information.
Configure
Apply product rules, materials, dimensions, PV assumptions, costs, manufacturing logic and project requirements.
Model
Clone a Working Reference Model for the project and connect current facts to relevant decision processes.
Evaluate
Assess direct and propagated evidence, uncertainty intervals, inconsistencies and critical blockers.
Clarify
Ask the next most valuable question and explain the current decision state in controlled language.
Review
Record human approval, the supporting evidence, model version, sources and the agreed next action.
First Industrial Showcase
Tensevo Configurator
Tensevo is a solar active lightweight system for large photovoltaic canopies. It provides a demanding real world test case because every project combines geometry, structural behaviour, photovoltaic yield, local conditions, technical standards, approvals, manufacturing, assembly, costs and customer objectives.
The purpose of the showcase is not to fund a single product configurator. It is to validate a reusable Decision Intelligence architecture under real industrial conditions.
Formfinder
Provides product neutral geometry, physical values, model provenance and the Universal Geometry Model.
Configurator Core
Adds location, materials, product rules, PV assumptions, prices, manufacturing logic and project data.
HPM
Represents decision readiness, necessary and sufficient relationships, evidence, uncertainty and blockers.
AQA
Targets the questions that can reduce the most relevant uncertainty or resolve a critical process.
LLM Layer
Explains the evidence state in accessible language without creating unsupported technical decisions.
Human Review
Maintains responsibility through named roles, explicit approval states and a traceable decision record.
Technology Architecture
A product independent decision layer
The Strategyfinder Configurator separates the reusable decision technology from the specific product logic of each industrial portal.
1
Geometry and Data Layer
Neutral geometry, site capture, parameters, physical values, loads, supports, versions and model relationships are exchanged through a shared data structure.
2
Product Logic Layer
Each product portal contributes its own rules, materials, pricing, interfaces, manufacturing constraints, documentation and commercial workflow.
3
Decision Intelligence Layer
HPM, evidence connectors, adaptive questions, controlled explanations, decision traces and review workflows provide the reusable platform core.
Working Reference Model
A reusable reference model is cloned for every project. Each clone contains project specific evidence, deviations, assumptions, open issues and learning points while the reference model remains version controlled.
Evidence Bound Explanation
Every relevant explanation can be connected to a process node, evidence status, source, model version and human review state. The language model acts as an explanation layer, not as an approval authority.
Human Control by Design
Not an automated approval machine
Strategyfinder prepares decisions. It does not replace engineers, planners, authorities, manufacturers or management responsibility.
Evidence remains visible
Engineering values, documents, assumptions, expert judgements and measurements are not treated as equivalent without defined translation rules.
Uncertainty remains explicit
The platform does not hide incomplete knowledge behind a simple status light. It shows where the evidence interval remains wide or inconsistent.
Critical blockers are surfaced
Missing site approval, unclear responsibilities, unsupported load assumptions or an unapproved product variant can be identified before they create expensive rework.
Final approval remains human
Every final release records the responsible role, evidence status, time, model version and decision comment.
AWS Development Focus
Research, development and measurable validation
The development programme focuses on the technical questions that must be solved before the platform can scale beyond the first showcase.
Core Development Questions
- How can geometry and physical values be translated into decision relevant process semantics?
- How can heterogeneous evidence from models, documents, measurements and expert judgement be evaluated consistently?
- How can uncertainty be reduced and documented over the project lifecycle?
- How can HPM guide adaptive questions toward the highest decision value?
- How can language model explanations remain bound to evidence, sources and review states?
- How can the architecture transfer from Tensevo to a second engineered product class?
Measurable Success Criteria
- Traceable recommendations with evidence, source, model version and approval status.
- Reduction of uncertainty intervals during project development.
- Earlier detection of critical blockers and inconsistent assumptions.
- Higher relevance of questions compared with static questionnaires.
- More consistent decisions across comparable project cases.
- Reduced rework and clearer responsibility for the next action.
Scalability
Designed to move beyond one product
Tensevo is the first validation environment, not the limit of the platform. The reusable technology consists of the reference model logic, evidence connectors, adaptive question policy, controlled explanation layer, decision trace and human review workflow.
Reusable Core
The same decision architecture can support different engineered product portals while preserving clear boundaries between generic platform functions and product specific rules.
Enterprise Integration
Manufacturers can connect their existing product data, engineering services, configurators and project workflows without transferring ownership of their customer projects.
Organisational Learning
Project clones create a structured basis for learning which evidence, questions and review points improve decision quality across future projects.
Confidential Showcase
Explore the first industrial validation environment
The Tensevo Configurator demonstrates how Strategyfinder connects product configuration, engineering evidence, uncertainty, controlled explanation and accountable human decision making in one integrated environment.