Built for your security team's questions
Before C16 connects to your SAP, your security team needs to know exactly what it can and cannot do. This page sets it out in one place. The full detail is in our security pack, shared under NDA.
What C16 can and cannot do in your SAP
C16 never writes to your production system and never changes your business data. No documents posted, no master data edited, no transactions created.
When you ask for code, it is written in your development system, your experts approve it, and it moves forward through your own change process. Nothing is promoted on its own. You choose the system C16 connects to.
Access and authorizations
- Each C16 user works through a paired SAP user that you create and control.
- We provide a least-privilege role template. Your team reviews it and assigns it.
- C16 cannot do anything that SAP user is not allowed to do. Your SAP authorization model is the boundary.
- Access can be withdrawn at any time by locking or removing the SAP user.
Connectivity and hosting
C16 reaches your SAP over a channel your team sets up and controls: a site-to-site VPN, AWS PrivateLink, or SAP BTP with SAP Cloud Connector. You keep control of firewall and network access throughout.
For strict data residency or isolation requirements, a dedicated single-tenant deployment in your own AWS account can be evaluated on request. It confines the platform, the SAP connectivity and all C16-side storage to your AWS boundary. AI inference runs as an external service; the data-processing and cross-border terms are set out in the Data Processing Agreement.
Your data
- No customer prompts, code or data are used to train any AI model, and nothing is shared with other customers.
- Prompts, responses and the SAP data returned during a conversation are stored with that conversation in your C16 account, so you can revisit and audit them.
- Generated documents are kept until you delete them.
- All customer data is removed on request, or within 30 days of contract end under our contractual commitment.
The AI model
C16 uses Anthropic's Claude as its underlying language model, standardised across all deployments so quality, behaviour and security controls stay consistent for every customer. Under the model provider's commercial terms, prompt inputs and outputs are not used for model training and are kept by the provider only for a limited period for safety and abuse monitoring.
Encryption and logging
- Data is encrypted in transit and at rest.
- SAP credentials are encrypted at field level with AES-256-GCM. Platform secrets are held in AWS Secrets Manager.
- Audit events are logged to AWS CloudWatch and can be forwarded to your SIEM.
- Log redaction automatically strips passwords and API keys.
Code changes and traceability
Every deployment needs explicit approval in C16 and goes through the standard SAP transport system, so SAP keeps its own trail: transport ID, reviewer, timestamp and change log. Your release team controls every transport onward. C16 keeps an independent audit log of user actions, available on request.
There is no lock-in. Generated code is standard ABAP, versioned by your transport system, and you own it.
Assurance
PEOL Technologies, which builds C16, is ISO 27001 certified and SOC 2 Type II compliant, with security controls aligned to ISO 27001 Annex A.8.24. We conduct periodic external penetration testing.
Talk to us about security
A 30-minute session for your security team: how C16 connects to your SAP, what it can and cannot do, and how your data is handled. Bring your questions.
After the session, we can share our detailed security pack under NDA:
- SOC 2 Type II report and ISO 27001 certificate
- Architecture and data-flow overview
- Connectivity prerequisites and the SAP role specification
- Data Processing Agreement, including the list of sub-processors
- Answers to your security questionnaire
More questions on how C16 works? See the FAQ.