General AI tools do not know your processes or how your business objects relate. Shugyo supplies that context through an ontology linked to your company data, then uses the model to answer operational questions.
A general assistant can retrieve text, but it cannot infer the definitions and relationships specific to your company. Without a shared model, it does not know:
Shugyo gives the AI an approved organization model.
"I don't have access to your invoice data. Common causes include approval and payment delays."
"23 invoices worth $182K are past the 5-day verification SLA. Most are assigned to two reviewers."
An agent with plugins searches each connected system when a question arrives. That can handle a lookup in one source. Questions across systems also need agreed object definitions and relationships.
An ontology describes the objects in your company, their attributes, and the relationships between them.
| Element | Description | Example |
|---|---|---|
| Objects | Things that exist | Invoice, Client, Employee, Project |
| Attributes | Properties of objects | Amount, status, due date, owner |
| Relationships | How objects connect | Invoice belongs to Client |
| Flows | How objects move | Draft > Verified > Sent > Paid |
Start with leader context, business terms, relationships, and authority. When a structured warehouse path is useful, Shugyo can map BigQuery tables and typed JOINs into the approved ontology.
A compact ontology snapshot goes into the agent's system prompt. It guides the SQL the agent writes and keeps queries within your approved model.
The ontology gives your data and the AI the same vocabulary. A question about "delayed invoices" uses your definition of an invoice and its due date.
Each question uses the same approved model, so a metric keeps the same definition across answers.
The ontology is the approved model of the business. The digital twin connects that model to useful evidence. In BigQuery deployments, entities can map to tables and relationships to typed JOINs that the AI can traverse.
The twin records what happened in the source systems.
Ask a plain-language question about your connected data.
Shugyo writes the query and returns the supporting records.
Flag process times and patterns that depart from an approved baseline.
Alerts identify the process, baseline, and size of the deviation.
Track how an approved metric changes over a selected period.
The trend uses the same metric definition for each period.
AI suggests process changes based on measured bottlenecks.
Review the evidence and projected impact before approving a change.
The ontology maps the terms in your question to the company model. Shugyo queries the corresponding data in the digital twin and returns the answer with its trace.
Each answer links back to source data.
Shugyo exposes the evidence and calculation behind an answer. You can inspect:
Review the evidence and record your feedback.
AI can propose an answer or a process change. A person reviews the source evidence and decides whether to act.
Each conclusion links to its source data and query.
A person must approve critical decisions proposed by the AI.
Ask why to inspect the logic behind a result.
Named owners review AI-supported decisions and monitor their outcomes.
You retain control of the company data used by the model.
You retain ownership of your data.
We do not use your data to train models.
Configure which users access which data. Role-based permissions.
Data encrypted at rest and in transit.
SOC 2 Type II in progress. AI Act ready.
Bring an operational question and the systems that hold the relevant data. We will show how Shugyo models the terms, runs the query, and exposes the supporting records.