Spend the time you spent searching on work
No more wondering which system holds it or whom to ask. One search finds it.
ECHO Catalog
ECHO Catalog is the registry of your company's AI. Register data, AI models, and business connectors, and the Orchestrator and Analyze work from this list. Data stays right where it is.
- Data · AI models · connectors in one list
- 3 types
- Data moved just to register it
- 0
- Permissions and masking by department
- Row · value level
- AI usage log
- Per call
Key benefits
Register once, and every AI uses it
No separate AI for each department. The Orchestrator, Analyze, and business apps share the registered assets.
Answer audits and security reviews instantly
Download the record of when AI used which data and submit it as is.
Keep going even when owners change
Descriptions, owners, and usage logs are attached to each asset, so handovers are short.
Who uses it
Connect the people who create data with those who use it
Data owners
Register descriptions and sources for tables and sensor data, organizing info so analysts can choose what to use.
AI developers
Search the purpose of models and connectors, read what goes in and comes out, and use them to compose AI tasks.
IT operators
Assign asset owners and usage permissions, and review usage to keep the asset list current.
Functions
Find it, understand it, trust it
01
Find it instantly by search
Search by name, type, department, or tag, and preview descriptions and sample values before choosing.
02
Same words, same meaning
Define terms each department used differently, like yield and uptime, in one glossary. AI answers with the same meaning.
03
Trace where it came from
Lineage shows the path data took, from source systems through rollups and AI models to reports.
04
See data health as a score
Scores missing values, duplicates, and last refresh time, so you know up front whether data can be trusted.
05
Register business functions as AI tools
Describe the inputs and outputs of functions like work order lookup or sending email, and AI reads the descriptions to pick the right one.
06
Mask at the row and value level
Rows visible are split by department and title, with unit prices · personal data masked. AI answers follow the same rules.
Screens
See your AI assets on one screen
Screen 1 · search, data name, source, type, owning team
Find the data you need instantly
"Where is the equipment vibration data?" No more hunting for someone to ask. Search, and you instantly see where it came from and who owns it.
- Search by name
- Check source
- Straight to the owner
Screen 2 · allowed, masked, blocked
Payroll · cost hidden even from AI
Define the data AI may use for each task. Cost unit prices are used masked, and HR records are blocked entirely.
- Allowed
- Masked
- Blocked
Screen 3 · time, AI task, data · functions used, result
Submit AI usage logs for audits
Every use of AI, when and what, is recorded and ready to show as is for audits and security reviews.
- Log by time
- Awaiting approval · blocked shown
- Submit as is
Use cases
When Catalog has the answer
When report numbers differ by department
When Production and Quality report different yields, open the glossary and lineage to see what basis and source each used, and align on one.
When a sales rep asks AI about cost
AI answers but masks unit prices outside your permissions. The same question shows a different scope depending on who asks.
When source data stops
If updates stop, the health score drops, and owners of AI models and reports that use that data are notified first.
When building a new AI task
Before building from scratch, search for models and connectors that already serve the same purpose, and reuse them.
Use-case scenarios
The foundation for the Orchestrator and Analyze
Add products one at a time as needed, and expand like this.
- LINKER Organize equipment data items.
- Catalog Register sources and owners.
- Analyze Analysts choose training data.
- Analyze Train and evaluate models.
- Catalog Register model purpose and call info.
- Orchestrator AI uses the model.
- ERP Configure ERP lookup tools.
- Catalog Describe inputs, outputs, and usage permissions.
- Orchestrator AI queries order data.
Alongside existing systems
Don't move assets. Just register them
Deployment model
Configuration and cost vary with scope, user count, and installation environment. Tell us your situation and we will recommend the right approach.
Contact salesInstallation environment
Design the registration scheme around asset locations and access boundaries: on-premises, cloud, or internal network.
Existing system integration
Links business system data and tools with source and owner info.
Open standards
Connector inputs and output descriptions are organized as asset info. Data and models use each system's identifiers and call paths.
FAQ
FAQ
Why register separately?
To keep AI from using just any data, decide first what it may use. Since it uses only what is registered, the source of every result is clear.
Do I have to move data into the Catalog?
No. Data stays where it is; only name · description · source · owner are registered.
Do I change AI settings here too?
Yes. Change AI models, instructions, connectors, and task-plan settings in the Catalog, and the Orchestrator picks them up immediately.
How is it controlled technically?
Register MCP tool endpoints and parameter schemas, and apply allowlists, role- and attribute-based permissions (RBAC · ABAC), row filters and masking, and per-call audit logs.