Overview
Data Agents enable you to automatically enrich and continuously update your application data in Torii using AI-powered queries.
With Data Agents, you can define your own AI prompts to collect the exact information your organization needs about vendors and applications — such as security posture, compliance certifications, risk indicators, public reviews, acquisitions, or recent breaches. Each field is automatically populated and refreshed based on a defined schedule, using your organization’s own OpenAI API key.
This elevates Torii into a continuously enriched SaaS intelligence platform, enabling faster decisions, stronger governance, and more impactful automation across your organization.
Benefits
1. Enrich and Expand Your SaaS Intelligence
Data Agents allow you to continuously enrich your application data with the exact insights your organization wants to monitor.
Instead of relying solely on static vendor information, you can dynamically layer AI-powered intelligence on top of your SaaS stack — tailored to your governance model, risk framework, and procurement strategy.
Example insights you can automatically generate and populate:
Has this vendor experienced a recent security breach?
Does this vendor host data in the EU?
What is this vendor’s current G2 rating?
Does this vendor comply with specific regulations (e.g., FDA 21 CFR Part 820)?
Has this vendor undergone a recent acquisition or major organizational change?
You define what matters. Data Agents ensure that insight is automatically populated and kept up to date — turning your application inventory into a continuously enriched source of vendor intelligence.
2. Keep Data Fresh Automatically
Data Agents run on a defined schedule (e.g., daily, weekly, monthly), ensuring your data remains up to date without manual research.
3. Power Governance & Automation
Once populated, Data Agent fields behave like any other application field in Torii. You can:
Use them to filter and sort the Applications table
Use them in Dashboards
Trigger Workflows
Use them as pesonalization tokens in workflows
4. Maintain Cost Transparency
Smart Fields use your organization’s own OpenAI API key. AI usage is billed directly by OpenAI to your organization, giving you full visibility and control over usage and cost.
How It Works
Data Agents are a new type of application group that runs a structured AI query against selected apps.
Each Data Agent contains a single AI query, can apply to specific apps or filtered groups of apps, and can populate multiple structured fields from one AI response.
Step 1: Configure OpenAI API Key
Before creating Data Agents:
Go to Settings → AI Settings
Enter your organization’s OpenAI API key
(Required to enable GPT-based features including Data Agents)
Step 2: Create a Data Agent
To create a data agent:
Go to Settings -> Application Details
Click Add data agent
Name the Data Agent
Start by giving your Data Agent a meaningful name based on its purpose — for example, “Security Overview” or “Review Platform Ratings.”
2. Enter AI Query.
Define the AI query that will be sent to Open AI.
For example - "For the current vendor provide current ratings on G2, Capterra, and PeerSpot"
3. Set Applicable Applications.
Define whether the query should run for:
All applications
Specific apps based on filter criteria (e.g., sanctioned apps only)
4. Generate Sample Response.
Click Generate to preview a sample response based on one matching application.
If you adjust your query, simply regenerate the sample to reflect the changes.
5. Create Output Fields
You can manually create the output fields, or use the Auto Generate option to automatically create structured fields based on the AI response (recommended).
Use the preview to confirm that the fields populate as expected, and adjust names or field types if needed.
6. Set Update Frequency
Choose how frequently the Data Agent should run.
Update frequency directly impacts usage cost, and an estimated cost is displayed during configuration based on the number of apps and selected cadence.
Once saved, Data Agent fields may take up to 1 hour to populate across the platform.
Step 3: Use AI Agent Output
Once created, Data Agent fields are available across Torii:
-
Applications page.
On the Applications page, each field appears as its own column, which can be filtered and sorted like any other field.
-
Application info tab.
Within the Application Info tab, the fields are updated automatically based on your schedule. You can review change history to see each update iteration. These fields are AI-generated and are not editable.
3. Workflows
The fields will be automatically available as personalization tokens.
They can also be selected as filter in Application meets criteria trigger configuration.
4. Dashboards
The fields are available in widgets once populated.
Best Practices
1. Start with High-Impact Use Cases
Focus first on governance-critical insights:
Security posture
Compliance certifications
Data residency
Risk indicators
Recent incidents
Specific compliance regulations
2. Structure Queries Clearly
When writing AI queries, be clear and structured. The more specific your instructions, the more predictable and usable the output will be. Using Auto Generate to create structured fields from the sample response is strongly recommended.
3. Use Data Agents Strategically
Consider grouping related insights into a single Data Agent. For example, a “Security Overview” agent might populate breach history, certifications, and risk classification from one structured query.
4. Leverage Data Agents in Workflows
Use data agents to:
Notify SecOps when compliance status changes
Get updates about recent security breaches
5. Monitor AI Usage & Scheduling
Align update frequency with how dynamic the information is. Highly sensitive risk indicators may require more frequent updates, while more stable data may not.
6. Hide (Don’t Delete) When Testing
When testing, hide fields instead of deleting them if you want to temporarily remove them from view. The data will continue updating in the background.
FAQs
Q: Who pays for AI usage?
A: Your organization does. Data Agents use your own OpenAI API key, and usage is billed directly by OpenAI.
Q: What happens if an AI run fails?
A: If the field has never been populated, it remains empty. If it was previously populated, the last successful value is retained .
Q: Are queries run only against the model, or is web search involved?
A: Data Agents use the OpenAI model and include web search capabilities to retrieve up-to-date publicly available information .
Q: Can I choose a different LLM or model?
A: Not currently. Data Agents currently use GPT-4o .
Q: What happens if I run out of OpenAI tokens?
A: If your token limit is reached, fields within data agent will not update. Existing values will remain unchanged .
Q: How quickly are fields populated after creation?
A: When you create a new Data Agent, it runs immediately and fields populate shortly after.
Q: Can I manually edit AI populated field values?
No. These fields are automatically populated and are not editable from the Application Info tab .
Q: Can I use existing Torii data inside Data Agent queries?
Not currently. Data Agents currently operate based on the defined AI query and publicly available data.