Business software should reduce operating cost—not create a new layer of cost, complexity and dependency.
We design bespoke applications around the way an organisation actually works, then engineer them to run efficiently on compact, right-sized infrastructure. The result is lower operational friction, lower lifetime technology cost and greater control over data, deployment and supplier relationships.
Bespoke operational applications designed around the way the organisation actually creates value.
Lean, micro-server-native systems that run efficiently on compact, right-sized infrastructure.
SOLID-aligned, data-sovereign architecture so data, identity and permissions remain portable and under the organisation's control.
Recurring data entry, spreadsheet reconciliation, disconnected subscriptions, manual approvals and informal workarounds are a continuous operating cost that rarely appears in the software purchase decision.
Bespoke Operational Applications →Most operational systems do not require hyperscale infrastructure. A well-designed application engineered for a micro-server costs less to run, requires less specialist support and introduces fewer dependencies.
Micro-Server Architecture →SOLID-aligned design keeps data portable, permissions explicit and hosting independent. Supplier leverage, exit cost and long-term options are preserved.
Data Sovereignty →The causal chain that connects operational design to infrastructure cost:
Bespoke software removes operational waste. Micro-server architecture removes infrastructure waste. SOLID-aligned design removes strategic dependency.
The engineering work below is the practical support for the business case: smaller infrastructure, more efficient processes, trustworthy data and useful AI.
01
Many business applications are over-infrastructured. Better application design, caching, asynchronous processing, modular services and restrained dependencies can materially reduce cost, complexity, energy use and waste.
Design the workload properly before buying more infrastructure.
Explore the architecture →02
AI does not have to run exclusively in hyperscale data centres. Local, small-node, federated and hybrid inference can route work according to capability, privacy, latency and cost.
Use central AI where it adds value, not merely by default.
See distributed AI →03
Businesses need inexpensive access to cryptographic trust. Digital signatures, Merkle logs, verifiable credentials and selective blockchain anchoring can provide integrity, provenance and independently verifiable records without placing confidential data on-chain.
Blockchain is a trust mechanism, not a place to store everything.
See the trust model →04
AI-assisted processes should operate within defined permissions, execution rules, validation gates and human escalation points. Automation should be controlled by the business process rather than by unconstrained model output.
Automate deliberately, with authority and accountability made explicit.
See AI governance →05
Token consumption is a technical cost metric, not a measure of business value. AI should be assessed through quality, risk, time saved, cost avoided, revenue contribution and net business value delivered.
Businesses should not issue blank cheques based on model consumption.
Measure outcomes →Full ERP Systems
Complete ERP solutions built for lean infrastructure. Optimised for single-node micro-server deployment. Our platform: VentES.
Big Data & AI
Analytics and ML pipelines designed for local, private and hybrid inference. Workloads are routed by privacy, capability, latency and cost rather than defaulting to a central model.
Blockchain & E-commerce
Selective cryptographic trust for provenance, auditability and commerce. Sensitive workflows stay off-chain and the trust model stays explicit.
VentES by kawalek.uk
We build systems aligned with Tim Berners-Lee's Solid vision — decentralised, privacy-respecting, user-controlled.
These are not architectural preferences for their own sake. Each principle protects a commercial interest: operating cost, business continuity, supplier choice, regulatory evidence, data value or future strategic options.
Applications Do Not Own Data — They Access It
Ownership lives with the user. The application is a lens, not a vault.
Solid-alignedIdentity Must Be Portable and Federated
Authentication uses open standards from the start. Identity is not owned by the application.
FederatedSystems Should Be Eco-Efficient by Design
Right-sized infrastructure. Pre-generated artefacts over repeated compute. Small-server deployment is a first-class topology.
Eco-efficientMost operational applications do not serve millions of simultaneous users. They should not automatically inherit the cost, complexity and dependency of systems designed for global hyperscale.
We design the application to run on a compact, right-sized server. Public cloud and managed services are introduced selectively — only where a workload genuinely requires elasticity, geographic distribution or specialised capability.
Cloud-capable, but not cloud-captive.
Minimal footprint
Only essential state on the app server. Storage stratified by data type.
Lean request path
Expensive work in Celery task queues. Synchronous path stays compact.
Async-first
Report generation, file processing, notifications — all deferred.
Eco-by-design
Right-sized infrastructure reduces energy, cost and environmental impact.
AI does not have to run exclusively in hyperscale data centres. Local, small-node, federated and hybrid inference can route work according to capability, privacy, cost, latency and energy use.
Tokens are a computation and cost unit. They do not prove useful work. AI execution needs rules and controls, and outcomes should be measured through quality, risk, time saved, cost avoided and net business value.
Tokens measure computation. Outcomes measure value.
We use the lightest mechanism that can be independently verified. The goal is integrity and provenance, not putting every business record on a public chain.
Use the right primitives for the trust problem
Digital signatures, hashes, append-only or Merkle-backed logs, verifiable credentials and external anchoring can provide integrity, provenance and independently verifiable records without exposing confidential business data on-chain.
Confidential business data should not normally be placed on a public chain.
Commerce stays supported
E-commerce remains an application example, but it sits behind the trust model rather than defining it.
We design and build our own micro-servers. Clients get a holistic, fully controlled, eco-conscious stack.
The SME Digital Adoption Taskforce estimates that a 1% productivity uplift across the UK's 5.5 million SMEs could add £94 billion annually to GDP. Cost, perceived difficulty and switching risk are identified as the main barriers to adoption.
Published evidence informs our commercial reasoning. It does not constitute a guarantee of project results.
Read evidence and sources →£94bn
Estimated annual GDP uplift from a 1% SME productivity improvement across 5.5 million UK businesses.
UK SME Digital Adoption Taskforce — final report, 31 July 2025
Tell us what you want to build and we'll get back to you directly.
