TIO AUTOPLAT · Application layer · Desktop + Web

AIDAPrivate enterprise agent

An agent deployed in your own environment, available as a desktop app and on the web. Ask data, query knowledge and have digital employees get work done in one conversation, with results pinned to ABI dashboards. Offline delivery and standalone operation; connect DataHub and AIHub and the same pages get richer instead of being rebuilt.

2
clients: desktop app + web, one backend
Offline
installs without internet access
Standalone
ask data, dashboards and knowledge without a platform
4
conversation modes, routed automatically

The agent lives in your own environment

Desktop and web clients talk to the same aida-server; data, metrics, knowledge and model calls stay inside your boundary.

Enterprise network · private deploymentDesktop appWebaida-serverBusiness dataKnowledgeModel servicePublic cloud
Desktop and web
The desktop app and web share one business API; WeChat adds push and light actions.
Offline delivery
Images and dependencies ship in the package; installation never pulls from the internet.
Local permissions and SSO
Organizations, roles and data scopes are managed locally, with standard OIDC single sign-on.
One definition per metric
Each metric has one editable definition; with DataHub connected, governed definitions win.

Same word "agent", different place to live

AIDA is for teams that must keep data and processes inside the enterprise.

DimensionPublic cloud agentAIDA private enterprise agent
Where it runsVendor cloudYour own environment, installable offline
ClientsWeb page or chat windowDesktop app + web, WeChat as a supplement
MetricsThe model guesses fieldsSemantic model and governed metrics, permissions injected
WritesDepends on the platformPreview, confirm, audit
Enterprise platformsSeparate integration workRicher in place once DataHub / AIHub connect

One conversation, four modes

AUTO only routes the business path — it never picks a model for you. The three business modes can also be chosen manually.

AUTO
Auto
Understands intent and routes to data, a digital employee or knowledge.
“East region sales last month? → Ask data”
ASK_DATA
Ask data
Maps to metrics and dimensions in the semantic model, generates a governed query and injects permissions.
“Split month over month by product line”
AGENT
Digital employee
Continue the same conversation with a chosen role; write actions are previewed, confirmed and audited.
“Have the dealer assistant send restock suggestions”
KNOWLEDGE
Knowledge
Answers come with source passages; no source, no claim.
“When does the new rebate policy take effect?”

Agent and digital employees

Published roles are ready here: pick an employee and you continue the same conversation, not a separate orchestration tool.

01
Employee list
See published digital employees and their status, and start working in one click.
02
One conversation
Hand a data conclusion straight to a digital employee to act on.
03
Gated writes
Changes to systems are always previewed, confirmed and audited.
04
To-dos and notices
Actions awaiting your confirmation are collected so nothing is missed.

General and industry roles

General rolesConfigured by formula
Customer service, tax, HR and admin roles configured in AIHub with one formula and used in AIDA.
Industry rolesProject-built
Inspectors, process engineers, dealer operations and content specialists, built with industry knowledge and systems.
Roles in productionLive in customer projects
Content generation and distribution and dealer operations roles are running in manufacturing customer projects.

Data questions, governed by a semantic model

Physical SQL is never exposed to users, and models never assemble SQL freely. Every number can explain its definition, scope and permissions.

01
Ask in plain language
“East region sales last month, month over month?”
02
Map to the semantic model
Metric: sales · Dimensions: region, product line · Time: last month
03
Governed query + permissions
Query generated from the model, row-level permissions injected before execution
04
Results with definitions
Tables, charts and metric notes returned together
05
Save to ABI
One click into cards, reports and dashboards
Model as contract
Metric definitions, default dimensions and drill levels are registered explicitly and versioned.
Aligned with DataHub
Import governed metrics from DataHub to avoid a second set of definitions in the application layer.
Permissions just work
Row and column policies are injected before execution, so everyone sees only their own scope.

ABI: conclusions from a conversation become reusable dashboards

Tables, charts and metrics from a conversation are saved through governed templates as cards, reports and dashboards; the cockpit is the full-screen view of the same dashboard.

Sales trend
Product mix
Region ranking
Inventory turnover
Collections
Alerts

Who works here

Analysts and department heads
Ask governed questions and turn tables, charts and metrics into dashboards.
Front-line roles
Enter the same conversation from the employee list; writes are previewed before confirmation.
Data administrators
Register metrics and dimensions in the semantic model, reusing DataHub definitions.
System administrators
Manage organizations, roles and data scope, shared with DataHub and AIHub.

In AUTOPLAT

AIDA is the agent people use every day; AIHub supplies AI and knowledge, DataHub supplies governed data.

Run it alone or as part of the full stack

Standalone
Deploy the portal, data questions, a lightweight knowledge base and semantic models with their own permissions.
Full stack
Connect DataHub to reuse governed definitions and row-level permissions; connect AIHub to use digital employees in the same conversation.

Deploy the agent in your own environment

Book a demo and see desktop and web clients handle data questions, knowledge and digital-employee work in one conversation.