Financial services
AI inference you can explain to your compliance team.
UK data residency. No US sub-processors. Full audit trail. No training on customer data.
FCA-regulated firms need to explain and evidence their AI supply chain. Marigold runs open-weight models on private AWS infrastructure in London. Your data does not leave the UK. No external model provider receives your inputs. Every inference job is logged within UK infrastructure.
FCA expectations on operational resilience and AI governance mean firms need to know where their AI runs, who controls it, and what happens to the data that passes through it. Marigold gives you a clear, documentable answer to all three.
Compliance-relevant properties
01
UK data residency
Inference runs on private AWS infrastructure in London. Data does not leave the UK as part of inference, storage, or queuing. No cross-border transfer occurs.
02
No US sub-processors
No inference request reaches OpenAI, Anthropic, Google, or any other external provider. The model runs locally from weights held on UK infrastructure. Your AI supply chain has one link.
03
No training on inputs
Customer data passed to the inference API is not retained for training, fine-tuning, or any secondary purpose. Outputs are stored briefly for retrieval and then deleted.
04
Audit trail
Every inference job records status, timing, and identity within UK infrastructure. Access logs are retained in CloudWatch within the UK region. Queryable by job and by time window.
05
Inspectable models
All models are open-weight with publicly documented architectures and training data. No proprietary black-box model receives your data. Model provenance is fully traceable.
06
UK-registered company
Bay Information Systems is a UK-registered company. Contractual relationships are with a UK entity subject to UK law and UK regulatory jurisdiction.
FCA and operational resilience
Know your AI supply chain
PS22/3 and the FCA's evolving expectations on AI governance require firms to understand and document the third parties in their operational infrastructure. An AI inference provider is a material third party if it processes customer data or supports a business service.
Marigold's architecture is designed to be simple to document: one provider, one region, no sub-processors for inference, full audit logs. We cannot assess whether our service meets your specific FCA obligations -- that depends on your use case, your important business services mapping, and your compliance team's judgement. What we can provide is the technical detail to support that assessment.
| Property | Marigold | US-based AI API |
|---|---|---|
| Data processed in UK | Yes | No |
| UK-registered processor | Yes | No |
| No US sub-processor for inference | Yes | No |
| No training on customer data | Yes | Varies / opt-out required |
| Audit log within UK region | Yes | No |
| Inspectable open-weight models | Yes | No |
Use cases
These are illustrative. Whether any specific use case is appropriate for your firm depends on the data involved and your compliance framework.
Customer correspondence classification
Classify incoming correspondence by type, urgency, and regulatory category. Route complaints, queries, and disclosures to the correct team. All processing within UK infrastructure.
Document summarisation and extraction
Summarise policy documents, contracts, and regulatory filings. Extract structured data from unstructured text. No document content leaves UK infrastructure.
Fraud and anomaly detection support
Generate embeddings for transaction narratives and customer communications. Use similarity search to surface anomalies against a known baseline. Private, auditable, UK-resident.
Frequently asked
Does Marigold process data outside the UK?
No. Inference runs on private AWS infrastructure in London (UK). Data does not leave that region. No US-based model provider receives your data.
Does Marigold train on customer data?
No. Inference inputs are not retained for training, fine-tuning, or any other purpose. Outputs are stored briefly for retrieval and then deleted.
Can Marigold support our FCA operational resilience obligations?
Marigold provides audit logs and UK-resident infrastructure with no US sub-processors. Whether this satisfies your firm's specific obligations under PS22/3 or related rules depends on your use case and your compliance team's assessment. We are happy to discuss the architecture in detail to support that process.
Is a data processing agreement available?
We are able to discuss data processing terms with firms that have specific requirements. We do not currently have a standard DPA document ready to issue. Contact us and we will work through your requirements directly.
What models are available?
The hosted registry includes instruct models for text generation and summarisation, embedding models for search and classification, and image-to-text models. All are open-weight with publicly documented provenance. See the full registry.
AI inference with a clear compliance story.
If you are evaluating AI tooling for a regulated firm, we would like to hear from you. Register your interest below or contact us directly.
Register interest
Tell us your firm type and use case and we will follow up directly.
Noted. We will be in touch.
Prefer to talk directly? ed@bayis.co.uk