HOW IT WORKS SHUGYO

From scattered data to operational clarity.

Three steps to visibility. A fourth that compounds. Weeks, not months - and no IT project to start it.

01 Week 1-2
Connect Systems connected
02 Week 3-4
Model Ontology defined
03 Week 5-6
Observe Twin goes live
04 Ongoing
Decide & review Run the rhythm
THE PROCESS01

Six weeks to insight. Then a cadence that compounds for years.

The first three steps give you visibility — your systems connected, your organization modelled, your digital twin live. The fourth is where the value compounds: leaders review the twin, make decisions, and watch how the organization responds, on a rhythm the business can sustain.

Visibility without rhythm is a dashboard nobody reads.

01-03

Visibility

Connect, model, observe. Your organization becomes queryable.

04

Compounding

Decide against the twin on a sustained cadence. Better decisions become measurable improvement over quarters.

STEP ONE · CONNECT02
STEP 01 Connect

Your existing systems. No migration required.

We connect to your data where it lives — ERP, CRM, databases, spreadsheets. There is nothing to move and nothing to rebuild.

We adapt to your data. You don't change how you work.

What we connect

Category
Examples
ERP & finance
SAP, NetSuite, QuickBooks, Xero
CRM & sales
Salesforce, HubSpot, Pipedrive
Databases
PostgreSQL, MySQL, SQL Server, MongoDB
Project management
Jira, Asana, Monday, ClickUp
Files
Google Sheets, Excel, Airtable
Custom
REST APIs, webhooks, custom databases

How connection works

01

Authenticate

OAuth or secure credentials. No code required.

02

Select data

Choose which tables, objects, or fields to sync.

03

Configure sync

Set the frequency: real-time CDC, hourly, or daily.

04

Validate

We verify data quality and completeness.

What you don't need to do

  • No data migration
  • No schema changes
  • No IT project
  • No consultants
STEP TWO · MODEL03
STEP 02 Model

AI proposes. You approve.

We build an ontology — a structured model of your organization. The AI does the heavy lifting. You bring the domain knowledge, and you decide what's right.

A formal description of your company — objects, attributes, and how they relate.

ontology · approved Example
ENTITIES
├─ Client │  ├─ name │  ├─ industry │  └─ status · active · churned
├─ Project │  ├─ name │  ├─ start_date │  └─ status · planning · active · done
├─ Invoice │  ├─ number │  ├─ amount │  └─ status · draft · sent · paid
└─ Employee    ├─ name    ├─ role    └─ department
Relationships Project → belongs to → Client Invoice → belongs to → Project Invoice → assigned to → Employee Employee → works on → Project

How modeling works

01

AI analyzes

Scans schemas, sample data, field names, and relationships.

02

AI proposes

Generates a draft ontology: objects, attributes, relationships, and flows.

03

You review

Rename objects, adjust relationships, and add or remove elements.

04

You approve

Once it reflects your reality, approve it. The digital twin builds automatically.

Why this matters

Shared understanding

Everyone agrees on definitions. No more "what do you mean by project?"

AI context

The ontology becomes the vocabulary AI uses to understand your questions.

Adaptable

As your organization changes, refine the ontology. The system adapts.

STEP THREE · OBSERVE04
STEP 03 Observe

Your digital twin goes live.

The ontology populates with real data. Your organization becomes queryable — ask in plain language and get answers grounded in the twin.

No SQL. No dashboard hunting. Just questions and answers.

AI assistant
Grounded
YOU
How many invoices are stuck in verification more than 5 days?
Context attached from ontology
Invoice Flow · verification Client SLA
23 invoices stuck — worth $182K
Traceable · invoices ⋈ flow_state · BigQuery

Ask questions

  • Which projects are over budget?
  • How long does client onboarding take on average?
  • Show me all invoices stuck in approval.

Track KPIs

Define metrics in business language. AI figures out how to calculate them.

01

Describe

Say what you want to measure, in business language.

02

AI proposes

The system proposes a calculation method.

03

See results

Run it on real data from the twin.

04

Approve or refine

Lock it in, or adjust until it is right.

Define processes

Create lenses — analytical perspectives on specific workflows.

Process
What you measure
Invoice processing
Time at each stage, bottlenecks, cost of delays
Client onboarding
Handoff delays, time-to-value, drop-off points
Sales to delivery
Pipeline accuracy, resource allocation
Monitor continuously

KPIs measured automatically - daily or weekly. Trends stay visible over time, so drift is caught before it becomes a crisis.

STEP FOUR · DECIDE & REVIEW05
STEP 04 Decide & review

Turn visibility into a decision rhythm.

A dashboard nobody reads changes nothing. The compounding gain is in the cadence. Decisions are made against the twin, recorded, then checked against their effects on the organization.

The operating loop is not "look at the data once." It is plan, do, check, act — at a frequency the business can sustain.

This is what turns better decisions into measurable improvement over quarters — not weeks.

01

Set the rhythm

Choose an evaluation frequency per process - daily, weekly, bi-weekly, monthly. Each process has an owner and a review cadence.

02

Decide on facts

Bring the twin into the room. End opinion-based debates with shared numbers.

03

Close the loop

Record what you decided and why. When the next review comes, you see whether it worked.

TIMELINE06

From signup to first insights.

Six to eight weeks to first insights. Not six to eight months. Then evolution never stops.

Week 1-2

Discovery

Understand systems, identify key processes, plan integrations.

Week 3-4

Connection

Connect data sources, validate quality, build the data layer.

Week 5-6

Modeling

AI proposes the ontology; you review and refine.

Week 7-8

Activation

Define initial processes, create KPIs, train users, go live.

Ongoing

Evolution

Add sources, refine the ontology, build new processes.

AFTER GO-LIVE07

The system grows with the business.

Add more data sources

Start with critical systems. Expand as needed.

Refine the ontology

As your organization changes, update the model. No migration required.

Build new processes

Define new lenses and KPIs as questions emerge.

Deepen AI usage

From queries to anomaly detection to predictive insights.

FAQ08

Common questions from evaluators.

How long does implementation take?

Six to eight weeks for first insights. We start with your most critical processes and expand from there.

Do we need to change our existing systems?

No. We connect to your systems as they are - no migration, no schema changes, no IT project.

What if our data is messy?

Most data is. We identify quality issues during connection, and the ontology layer normalizes the chaos into structure.

Who needs to be involved?

A sponsor, a domain expert, and someone who can grant data access. No dedicated IT team required.

What about security?

Data is encrypted at rest and in transit, access is role-based, and your data stays yours.

Can we start small?

Yes. Most customers start with two or three data sources and one or two processes, then expand as they see value.

What if the AI-proposed ontology is wrong?

You review and approve everything. AI proposes, you decide. Iterate until the model reflects your reality.

How is this different from BI tools?

BI tools show dashboards. Shugyo builds a living model you can query in natural language.

SEE IT ON YOUR DATA

Ready to see how it works for you?

Let's walk through how Shugyo would connect to your systems and model your organization - on your data, not a demo dataset.