SOSX®, AI-enabled systems thinking, for any sector.
For programme architects at defence and CNI primes
Programme Architect, Defence / CNI Prime
SOSX is AI-enabled systems thinking: an agentic AI platform for defining, researching and analysing systems-of-systems problems in any sector, with governance built in that makes it safe for enterprise.
A programme spanning dozens of subsystems and suppliers is a system of systems whether or not anything models it as one: the interactions exist; the question is whether anyone can see them. If you architect such programmes, the pains are familiar: system-of-systems interactions are opaque; assurance and governance evidence for AI-containing systems is bespoke every time; and supplier tooling rarely meets defence-grade trust expectations.
The discipline: how a programme is held together
You already run the practice; this is only us showing we know what it is. Requirements traced from need to verification, so nothing is satisfied by assertion. Interfaces controlled and owned, because the programme's real risk lives between subsystems rather than inside them. An assurance argument built as claims, argument and evidence, where each claim names what would falsify it. Configuration and baselines held, so "which version was reviewed?" has an answer. Supplier dependencies mapped with their lead times and their float, not just their contracts. Risk carried with pre- and post-mitigation severity and re-scored when the parameters move. And independent technical review (somebody outside the team, paid to disagree) before the review board rather than during it.
The hard part was never the method. It is that a programme spanning dozens of suppliers generates more cross-system consequence than any working group can hold in one place, and every answer assembled by hand is stale by the time it reaches the board. The expensive part is not drawing the model; it is building it and keeping it true.
The programme as a model, not a diagram
SOSX applies systems thinking to the programme itself: it maps subsystems, suppliers and interfaces as a typed, connected network with its reinforcing and balancing flows, parameterises it from cited sources, and runs grounded analyses across it: trade-off, what-if, risk, system dynamics.
Two properties matter most in this room. First, provable, not just plausible: an OWL-EL symbolic reasoner derives structural facts with the entailing axioms attached, kept visibly separate from LLM judgement and from computation, in an assurance argument, a reader always knows which kind of statement they are looking at. Second, it sits above the toolchain, not instead of it: SOSX exports into nine MBSE tools and formats (SysML v2, Cameo, Rhapsody, Sparx EA, Capella, Modelio, MATLAB/Simulink, ArchiMate, draw.io), and imports from the same nine, one way in each direction, so outputs land in the MBSE practice the programme already runs rather than in another silo.
The second network: your supply chain
A defence programme is two systems of systems, not one. The capability is a network, sensors, effectors, communications and command, delivering an effect only when connected. The supply chain underneath it is a network too: a multi-tier web of primes and smaller suppliers through which requirements, assurance obligations and engineering data have to flow.
Policy has caught up with that. The UK's industrial strategy moved defence away from buying on cost alone and toward treating critical industrial capability as a strategic asset, committing to map the most critical supply chains and manage their risk. National Audit Office scrutiny has made the same point from the other side: a supply chain optimised for cost is fragile when demand surges or supply is disrupted.
Both problems have the same shape, and SOSX treats them the same way: model the network, find the single points of failure and the hidden common dependencies, and run the disruption question before it arrives rather than after.
What flows down the tiers, and who holds it
Requirements and assurance obligations cascade down the sub-contracting tiers, which is how the MOD gains assurance over a chain it does not directly contract with, and AI-specific assurance expectations now add to that. The burden lands hardest where the tooling is thinnest: a tier-two or tier-three supplier is increasingly asked to produce its own assurance evidence without the heavyweight seat licences a prime takes for granted.
That is the gap SOSX is shaped for. Decision provenance, adversarial validation and an end-to-end audit trail give a smaller supplier an affordable, inspectable way to evidence assurance up the chain, and because SOSX exports into nine MBSE tools and formats, that supplier can take part in a prime's modelling practice without adopting its toolchain wholesale.
Trust markers, stated honestly
On trust, we'll say honestly what we can reference and what we can't. Our referenceable engagement is live enterprise MBSE trials with a large global aerospace manufacturer: trials, not production, and not a defence programme. We won't claim a reference we don't have, and your procurement process is precisely why.
The company behind SOSX: Digital Tactics, a 15-year UK SME (founded 2011, Hove) with Cyber Essentials certification and AWS partner status: more than 170 client projects, nine products. More on the about page; the engineering practice itself is SOSX Engineering.
The question you'd put to a review board
If a supplier slips a subsystem by six months, what does that do across the programme — and can I put the answer in front of a review board? With SOSX that is an auditable, cross-programme answer with its sources attached: in our experience, weeks and months of work in just a few hours.
Go deeper
Bring us one programme question
Bring us one programme question that normally takes a working group weeks (a supplier slip, an interface risk, a capability trade) and watch SOSX work it. We're in live enterprise trials, not production, and we'd rather earn the harder conversation with a worked answer than a claims deck.