We have developed to enable clients to effectively use AI.
Systems we have built and run, and the reference architecture we design from.
Systems we have built
Systems we offer as SaaS and systems we build for organizations. Real systems but Screens labeled as samples are illustrations with made-up data.
SaaSCorporate AI
An internal AI platform your people can use safely and your organization can control.
View details →
Built by usTodaytrace AI
An AI assistant for SOC teams that triages and investigates SIEM alerts, summarizes what happened, maps it to MITRE ATT&CK and suggests a response, with the analyst making the call.
View details →
Sample screenBuilt by usKnowledge Base & RAG
Answers from your own documents, with a source cited for every answer.
View details →
Sample screenBuilt by usAgent Team
A team of AI agents that splits up multi-step work, with people approving the steps that matter.
View details →Case studies
Case studies are on their way
We are getting client approval to publish our work. In the meantime, ask us for examples related to your problem.
Reference architecture
- BUILD KNOWLEDGE
- Sources: SharePoint / Drive; Documents, policies, approved exports; Data owner + classification
- Ingestion: Thai-aware parse + OCR; Validate / deduplicate / chunk / embed; Incremental sync + deletion propagation
- Knowledge & retrieval: Versioned content + metadata + ACLs; Hybrid search + rerank; Permission filtering before context
- SERVE + ACT
- Channels: Web / Chat / LINE / Teams; API gateway + SSO; Rate limits + user feedback
- Agent runtime: Orchestrator + model router; Context + session state; Grounded answers + source citations; Abstain or escalate if evidence is weak
- Tools & workflows: Live ERP / CRM / database APIs; Scoped access + validated inputs; Human approval for sensitive writes; Timeouts / retries / idempotency
Retrieve permitted evidence
SECURITY + GOVERNANCE: SSO / RBAC, PII controls, Encryption, Audit, Retention, Injection defenses
OPERATIONS + QUALITY: Traces, Retrieval + answer evals, Latency / cost budgets, CI/CD + rollback, Alerts + recovery
OWNERSHIP: Data owner, Platform, Security, AI engineering, Product
SUCCESS METRICS: Answer quality, Task success, p95 latency, Cost per task, Adoption
The knowledge and agent system structure we start from. It is a reference, not a specific client's system.
Find out how ready you are for AI
Answer 13 questions in about 5 minutes. You get a readiness score across 6 areas and a first scope to start from.