Workflow before tools
Understand how the work gets done before selecting a model, platform or vendor.
Systems thinking for AI that has to work in the real world
Codance helps organisations make better decisions about artificial intelligence before they commit to the wrong tool, supplier or build.
The consultancy begins with the real workflow: its purpose, information, decisions, exceptions, authority, consequences and recovery. AI is recommended only where it adds genuine value to a system that can support it.
Why Codance exists
Operational complexity does not begin at enterprise scale.
SMEs and mid-market organisations depend on the same fundamental kinds of operational information as the largest enterprises—customers, suppliers, enquiries, orders, jobs, documents, approvals, deadlines, costs and commitments.
They seldom have the same concentration of specialist staff, implementation capacity or tolerance for failed technology projects.
Codance makes rigorous workflow, data and AI-integration thinking available through proportionate engagements.
We reduce the size of the engagement, not the quality of the thinking.
Jonathan Arklay
Jonathan Arklay brings more than forty years of engineering, operational, design and solution-finding experience.
His background includes manufacturing and job-shop engineering, MRP integration, application logic, databases and data architecture, multidisciplinary engineering, technical management and systems analysis.
That experience established a durable principle: technology creates substantial value when it strengthens a workflow an organisation understands. When it is expected to impose order on operational confusion, the project often costs more, takes longer and disappoints the people relying on it.
AI makes the principle more important because it can produce a convincing answer even when the underlying information or instruction is ambiguous.
Jonathan’s role is therefore not to persuade every client to “adopt AI”. It is to help the organisation identify the right decision, boundary, control and first intervention.
Aurora: internal engineering and research
Aurora is Codance’s internal agentic-AI reference and R&D platform. It provides a practical environment for exploring:
- persistent, inspectable work state;
- PostgreSQL-centred operational truth;
- foreground and background workflow orchestration;
- model and provider abstraction;
- memory, embeddings and retrieval;
- secure secret handling;
- durable records and receipts;
- human authority, audit and recovery.
Aurora is described honestly as internal Codance engineering work. It is not presented as an external client deployment or used to invent client history.
This work helps keep the consultancy grounded in active engineering questions rather than recycled management language.
Codance principles
Deterministic work deserves deterministic software
AI should not replace a dependable rule merely because AI is fashionable.
People retain consequential authority
AI may assist, draft, classify or recommend. Named people remain accountable for important judgement and commitments.
Operational truth matters
Evidence, working status, proposed information and approved outcomes must not be silently blurred together.
Failure is part of the design
A credible integration explains how uncertainty is escalated and how work continues when a model, service or interface fails.
Independent judgement has value
“Not ready yet” or “AI is not the best first solution” can be the right and commercially valuable conclusion.
Care, clarity and conscience
The existing Codance holding page describes the practice as engineering systems with care, clarity and conscience. The AI consultancy retains that position:
- Care for the people, evidence and operational consequences involved.
- Clarity about what is known, assumed, authorised and excluded.
- Conscience about what should be built—not merely what can be built.
Company information
Codance Ltd is registered in England and Wales. Company number 17090892. Registered office: 15 Channings Drive, Exeter, United Kingdom, EX1 4AR.
Bring one real operational question
If you are deciding whether AI belongs in a workflow—or need a clear design before speaking to developers—Codance can help you make the decision on firmer ground.