The enterprise digital brain
Sense signals. Decide with rules and AI. Act in the moment. Learn from every correction — continuously.
Work email + one-time code. New here? You can register from the sign-in page.
- 4
- live adapters
- 4
- layer control loop
- 1
- shared core
Sense → Decide → Act → Learn, closed over a live Knowledge Fabric
What is EDB?
A reusable operating brain for live enterprise decisions
Most AI demos stop at a chat box or a static dashboard. Enterprise Digital Brain (EDB) is Trimurai’s Sense → Decide → Act → Learn scaffolding: a domain-agnostic control loop that turns live signals into governed actions, then gets smarter when operators correct it in plain English.
Domain apps only supply adapters and rules. The brain itself — event schema, DecisionObjects, Knowledge Fabric and the proactive agent — stays generic, so the same core powers finance, weather, maritime IoT and ITSM today, CRM and risk tomorrow.
Four live demos · one brain · many adapters
One brain, many domains
StockBuster, StormBuster, FleetPulse and OpsPulse are live today. The same Sense → Decide → Act → Learn core is designed for CRM, risk and more — without rewriting the loop.
Not another dashboard
EDB is an operating layer: it ingests normalized events, produces DecisionObjects, executes side effects, and remembers what operators taught it.
Governed autonomy
Rules and Gemini share the Decide layer. Humans stay in the loop via Learn — corrections become policy, not one-off dismissals.
Continuous improvement
Every exception, suppress rule and autonomous insight feeds the Knowledge Fabric so the next similar signal is smarter.
Architecture
Layered intelligence model
EDB is built as four stacked layers. Operators work at the top; the cognition loop reasons in the middle; the Knowledge Fabric grounds everything below it; and autonomy surfaces read from the same memory.
Every layer is domain-agnostic — adapters plug in, the stack stays the same across StockBuster, StormBuster, FleetPulse and OpsPulse.
Every layer is domain-agnostic — adapters plug in, the stack stays the same
Operator experience
Findings queue, decision trail, Explore chat and a teach surface — the only places a human touches the brain.
Cognition loop
Sense, Decide, Act and Learn run as one governed pipeline over a universal event schema.
Knowledge Fabric
Vectors, embeddings, learned operating rules and a document corpus — the semantic memory that grounds every decision.
Autonomy surfaces
A proactive agent hunts patterns unprompted; Explore lets leaders query the brain and promote answers into rules.
SDAL pipeline
Four layers. One closed loop.
Every domain instantiation — StockBuster, StormBuster, FleetPulse, OpsPulse and the rest — runs the same cycle.
- 01
Sense
Adapters normalize live signals into one universal event schema — finance, weather, maritime IoT, and ITSM live today; CRM and risk next.
- 02
Decide
Rules plus Gemini refine every event into a DecisionObject: intent, confidence, exception.
- 03
Act
Exceptions, feed updates, alerts and webhooks fire from the same decision contract.
- 04
Learn
Plain-English operator feedback becomes durable operating rules in the Knowledge Fabric.
Universal contract
Any event in. One DecisionObject out.
Adapters never talk to the UI directly. They normalize into the EDB event schema; the Decide layer returns a standard decision the Act layer can execute.
Domain-agnostic output — the same shape for every adapter
Knowledge Fabric
Memory, rules and recall — a shared substrate
The brain remembers what it decided, the policies operators taught it, similar patterns it can retrieve when the next signal arrives — and uploaded SOPs operators can search with citations (isolated from live Decide).
- Operating rules + decision traces (vector recall for exceptions)
- Document corpus Q&A — runbooks and SOPs per brain, search-only MVP
- SDAL trace replay — full Sense → Decide → Act path for any event
Learn layer
Teach it in plain English
Corrections become operating rules — not one-off dismissals. In StockBuster, teaching the brain to ignore banking volume spikes during options expiry persists for every similar signal.
The same Learn loop applies to proactive insights: dismiss noise, or confirm useful patterns to boost them on future agent scans. OpsPulse uses it the same way — for example, teaching the brain that SAP batch jobs are expected overnight noise. Export and import operating rules per brain for audit and portability.
Proactive agent
Insights before you ask
A background worker scans decision history and surfaces unprompted pattern insights — without waiting for a user query. Discovery runs in layers: vector anomaly and LLM clustering in the core brain, then domain group patterns in each live adapter (finance, weather, maritime, ITSM), with Gemini polishing a readable narrative alongside the technical detector output.
Run the agent on a schedule or on demand — the brain keeps thinking while leadership is elsewhere.
Explore & Set Intent
Ask first. Encode intent second.
When push insights and direct teaching are not enough, Explore lets executives investigate first — multi-turn chat over structured config (vessels or CIs), uploaded documents, and open findings. Live today in FleetPulse and OpsPulse. When sources disagree, EDB surfaces the conflict before you Set Intent and promote a pattern into a durable Learn rule.
Live demos
Four adapters. One core. All running now.
Live demo
FinanceStockBuster proves the brain on live markets
The finance adapter pulls an NSE watchlist, tags configurable sector groups, and runs a proactive agent you can trigger on demand — the same EDB core as the other live demos, isolated by brainId in Firestore.
- Live market ingest
- Configurable sector groups
- On-demand agent scans
Live demo
WeatherStormBuster proves the brain on live weather
Same SDAL loop, different adapter. StormBuster pulls Open-Meteo readings for eight Indian metros (including Pune), flags heat, humidity, wind and precipitation with Decide thresholds, and lets operators teach suppress rules in plain English.
- 24/7 weather ingest
- Regional agent patterns
- Learn overrides
Live demo
Maritime IoTFleetPulse proves the brain on voyage performance
Same SDAL loop, maritime adapter. FleetPulse ingests mock voyage telemetry (fuel, speed, engine RPM) for Indian Ocean trade lanes, flags overconsumption and speed deviation with Decide thresholds, supports document Q&A on SOPs, Explore & Set Intent for user-driven pattern discovery, and learns operator overrides in plain English.
- Voyage telemetry ingest
- Explore → Set Intent
- Document Q&A (RAG)
- Fleet agent patterns
Live demo
ITSM + CMDBOpsPulse proves the brain on incidents and change
Same SDAL loop, ServiceNow-shaped adapter. OpsPulse ingests ITSM readings for a demo CMDB, flags SLA risk, repeat incidents and change collisions, and lets you switch between a scripted Friday war-room and a UCI ServiceNow extract — then Explore & Set Intent.
- Friday + UCI data instances
- Explore → Set Intent
- Document Q&A (RAG)
- Service / site agent patterns
Get started
Bring the brain into your operating model
See how EDB fits your domain — or open any live demo with a work email.