SOSX®, AI-enabled systems thinking, for any sector.
System research, build and analysis
Run by us, on your material
You bring the question and what you already know about it. We do the rest: the research, the model, the analyses, and the outputs you can put in front of a board.
It suits a first serious question, the one where the honest answer is "it depends on about nine things" and nobody has the fortnight it would take by hand. It also suits a team who intend to run SOSX themselves but want the first network built properly, because the first one sets the shape of every one after it.
We work from your material and from the public record. Your briefings, your supplier filings, your cost models, whatever you are willing to share, alongside the literature, the regulations and the market evidence we go and find. Where the two disagree, you get told, because that disagreement is usually the most useful thing in the report.
What you actually get
The deliverable is not a slide with a recommendation on it. It is the reasoning, in a form you can interrogate and reuse.
- The research. The sources we found and read, catalogued, with every claim in the work citing one you can open and check.
- The model. Systems, sub-systems, components and actors as a typed, connected network, every value carrying its units, its provenance and its confidence.
- The analyses you asked for, each producing a grounded report.
- An explainability companion for every analysis, documenting the conversion chains, the assumptions and the confidence levels behind it.
- A per-system node report for any system in the network.
- The audit trail: every model interaction and document write logged, so the work can be reviewed rather than taken on trust.
- The model itself, exported into the tools you already run.
The analyses we can run
Each one produces its own grounded report: unit-aware parameters, citation-backed claims, uncertainty marked rather than smoothed away. Which of them earn their place depends on your question, and we would rather argue that out with you than sell you the whole suite.
Trade-off analysis
“Which option wins, and what does it cost us elsewhere?”
Multi-objective comparison across systems or scenarios, with the Pareto frontier drawn and the sensitivity caveats attached.
Compare models
“Which AI model does this work best?”
The same task on the same network, run by several AI backends, with time, cost, carbon and output quality set side by side.
Comparative study
“How do these options compare, fairly?”
Several variants built to one shared backbone, so the comparison is like for like rather than six differently-shaped models nobody can line up.
Product explorer
“What should this product actually be?”
A guided exploration of category, market, performance, constraints and design trade-offs, held inside the network.
What-if analysis
“What happens if we change this one thing?”
A hypothetical change written into a scenario copy of the network, and the impact traced through the feedback loops.
Deep analysis
“How does this system actually behave over time?”
System dynamics: the reinforcing and balancing loops, how they interact, the emergent properties no single part accounts for, and where the system is brittle.
Bayesian optimisation
“What settings give us the best outcome?”
Network parameters optimised, ranked by how much each one moves the objective, with the uncertainty quantified.
ROI analysis
“Does the money work, and when?”
Net present value, internal rate of return, payback and profitability index over the horizon you set.
Market analysis
“Who else is in this, and who decides?”
Demand, segmentation, the competitive landscape, who holds procurement authority, and where the lock-in sits.
KPI tracker
“What are we measuring, and who owns it?”
Performance measures defined against the model, with targets, trend and a named owner for each one.
Risk register
“What could go wrong, and how badly?”
Severity, occurrence and detection in the FMEA style, scored before and after the mitigations you propose.
SWOT
“Where are we strong, and where are we exposed?”
The familiar four quadrants, but built out of the model's loops and confidence levels rather than a workshop's memory.
Governance analysis
“Who decides this, and who has to be told?”
Decision rights, accountability and escalation paths, as a RACI matrix and the approval chains behind it.
Compliance analysis
“What does the audit trail show, framework by framework?”
SOSX's own audit evidence for a network, scored control by control against the governance frameworks it aligns with. Alignment, not certification.
Network quality critique
“How much should we trust this model?”
Centrality, single points of failure, orphans and coverage gaps, scored, with the fixes ranked by what they buy you.
Each of these has a page of its own, setting out what the analysis asks for and how SOSX runs it: browse the analysis tools.
What it lands in
Reports are produced as markdown and exported to PDF or Word, with headings, tables, inline formatting and embedded charts intact. The network itself exports into nine MBSE tools and formats (SysML v2, Cameo, Rhapsody, Sparx EA, Capella, Modelio, MATLAB/Simulink, ArchiMate, draw.io), alongside SOSX's own JSON, and an existing model in any of the same nine can be imported as the starting point, so the model lands in the practice you already run rather than in another silo.
Everything produced is indexed in a document library, so the pack you hand to a reviewer is a set of documents rather than a folder of screenshots.
Bring us the question
The engagement starts with a conversation about the question itself, because scoping it properly is most of the value and we would rather find out early if the answer is that you do not need us. Tell us what you are trying to work out and we will tell you what it would take.