Permitted uses
AI systems may assist development, analysis, documentation, automation, and engineering workflows—always subject to human review before customer delivery.
AI in delivery
AI usage policy
Where AI assists our work—and where human review, privacy-first practice, and your rules take precedence.
Synaptic Four uses AI-assisted systems pragmatically to support engineering and consulting workflows while maintaining human review and privacy-conscious operational practices.
Human oversight
We do not present raw model output as finished deliverables. Engineers validate logic, security implications, and fit with your constraints.
Privacy-first operating model
We prefer low-data-exposure workflows: avoiding direct access to customer personal data where the problem allows, and using anonymised or synthetic datasets for development.
Customer restrictions
You may request limits on specific tools, providers, or data flows. We document agreed restrictions and work within them.
Sensitive and regulated data
In consulting and delivery engagements, personal, health, or otherwise regulated customer data is not processed via third-party AI services by default. Additional safeguards and agreements apply when such processing is required. This is distinct from Synaptic Products apps you choose to use: SitePulse sends audio and derived text to named US AI/transcription providers as described in the Privacy Policy; optional BioResearch Assistant cloud LLMs run only if the operator configures their own keys. Kit generation on the operator side may use Anthropic, Gemini, Cerebras, or OpenAI. SynaptiSec and AZAV-Akte do not use LLMs.
Additional safeguards
On-premise models, air-gapped workflows, or no-AI clauses are supported when proportionate. EU AI Act–oriented technical reviews are offered separately where relevant.
This policy describes our standard practice for engagements. It is not legal advice and does not replace contract terms agreed in writing.
Related
Tell us your policies—we map a workable delivery setup.