ProductsJoyce
In developmentJoyce.Evidence interpreted. Law applied.
Joyce is Inveena’s neuro-symbolic legal and regulatory classification platform. It connects evidence to legal rules to establish which requirements apply.
Explore Joyce“In the particular
is contained
the universal.”
The cost of compliance
£33.9 billion.
Every year. The estimated cost of regulatory compliance across the UK financial services sector.1
TheCityUK & PwC · November 2025
Behind that figure: understanding which rules apply, determining what they require and building and maintaining the systems needed to comply.
What does the law
require of you?
Before an organisation can meet its obligations, it must establish which rules apply to its products, activities and circumstances. The answer depends on the evidence, the legal criteria and the context.
Repeated manual assessments add cost and operational complexity. Inconsistent classifications expose the business to rework, regulatory risk and decisions it may struggle to defend.
Introducing Joyce
From evidence
to determination.
Joyce brings AI interpretation and formal legal reasoning together. AI identifies relevant evidence; a governed reasoning engine applies the rules. This neuro-symbolic approach keeps material legal judgements under human expert oversight.
Interpret evidence
AI reads documents and supporting material to identify relevant facts, surface uncertainty and organise the evidence for review.
Apply legal rules
A governed symbolic reasoning engine applies formalised legal criteria to the established facts, using explicit, deterministic logic.
Explain the outcome
A determination connects the outcome to its rules and evidence. Human experts retain oversight of material judgements and unresolved questions.
Collateral opinion assessment
Existing opinions.
Automated analysis.
The Joyce Collateral Opinion application assesses an asset against existing ISDA collateral enforceability opinion frameworks, producing provisional answers and identifying where expert judgement is needed. The same approach is designed for other legal and regulatory domains.
Illustrative assessment
“Does the existing opinion framework cover this asset?”
A user uploads the asset’s legal terms, ownership, transfer and custody arrangements. A language model identifies the relevant facts.
A governed reasoning engine applies a machine-readable representation of the opinion’s questions, assumptions and qualifications.
WHERE the framework supports an answer
SHOW the evidence and reasoning
WHERE legal questions remain unresolved
FLAG the gap for expert judgement
Simplified illustration · not a live legal assessment
What this makes possible
Less uncertainty.
A clearer basis to act.
Designed to reduce the cost and complexity of classification, with evidence, rules and accountability in view.
Lower compliance costs.
Reduce repeated document review and manual classification work. Reuse formalised rules to simplify the cost of building and maintaining compliance systems.
More consistent decisions.
Apply the same governed criteria to comparable facts. Make gaps and conflicting evidence visible, supporting accuracy without concealing uncertainty.
A decision you can audit.
See which evidence and legal rules support each determination, giving reviewers a clear basis to scrutinise, challenge and explain the outcome.