Use the Insights Agent

Create reports, build dashboards, and get answers about your LogicGate data just by describing what you need, no manual report-building required.

The Insights Agent is a conversational way to work with your Risk Cloud data. Instead of manually configuring a report or dashboard, you describe what you want in plain English, and the agent builds it for you. You can also ask it questions directly about a specific record, a trend across your data, or a gap you want to check for, and it will answer using your organization's actual Risk Cloud data.

The Insights Agent is designed to be useful to any user, not just people who regularly build reports. Compliance analysts, system administrators, and executives can all use it to get to an answer faster than building a report from scratch.

Availability

The Insights Agent is currently available through an open beta program. It's under active development, so responses may occasionally be inaccurate and the agent may not always behave as expected.

Getting access

1. LogicGate enables the Insights Agent for your environment as part of enabling the open beta. If your organization wants to participate, talk to your LogicGate representative.
2. Once it's enabled, you can add the Insights Agent module entitlement to a role, either an existing role or a new one (for example, "Insights Agent Open Beta") and add the intended users to that role. This entitlement is what restricts access to a specific subset of users rather than automatically opening it to everyone specifically for the beta.
3. After being added, refresh the page. You should now see the AI button appear in your top navigation bar.

The agent uses the page you launched it from as context; for example, asking it a question from inside a specific application scopes its answer to that application's data.

How to use the Insights Agent

1. Click the AI button from any page.
2. Type your request in plain English, the same way you'd ask a colleague. You can ask it to build something, or just ask it a question.
3. The agent will display its thinking process as well as the underlying tool and agent calls it is doing in order to synthesize an answer.
4. Review what the agent returns. If it created or edited a report or dashboard, you can open and adjust it like any other report.

Your conversation stays active as you move between table reports, visual reports, and dashboards, so you can keep building on the same thread without starting over. 

You can also carry on a conversation across multiple prompts. For example:

You: What's our average time to remediate a critical finding in the Asset Management app? 

Agent: I found three workflows in this app: Vendors, Assets, and Documents. Would you like the average remediation time across all of them, or a specific workflow? 

You: Only the Vendors workflow. 

Agent: The average remediation time is 4 hours. 

You: Create a table report showing all critical findings with a remediation time above 4 hours. Include the owner and the date. 

Agent: Done — I've created a table report showing all findings with a remediation time above 4 hours, including the owner and date.

When the agent asks a clarifying question

If your prompt could reasonably be interpreted more than one way, the agent will pause and ask before proceeding rather than guess. This commonly happens when your wording could match more than one workflow or field, a term like "open" could map to more than one status value, a possessive like "my team" can't be confidently resolved, or an existing report looks similar enough to what you're asking for that it's unclear whether you want a new report or an edit to the existing one. Answer the question and the agent will continue from there.

You will notice sometimes the agent will create follow up action buttons for you. These are recommended next steps for you to take based on the conversation. You can easily click any of those follow ups or type another prompt if you have something else you want it to do.

What you can ask it to do

Build or edit reports and dashboards. Ask it to create or edit a single table report or visual report, or make the same change across several reports at once, including duplicating a report for each value of a field (for example, one copy per department) or swapping a field across every report that uses it. You can also ask it to adjust report settings like including archived records or enabling in-table editing, list your existing reports or dashboards, or recommend which reports to build (or add to a dashboard) based on the kind of insight you're looking for. For example:

• "Create a table report using the Business Processes workflow with the fields Business Process Name, Process Status, and Impact Score."
• "Show me the open issues across all applications, sorted by most important."
• "Give me a summary of this visual report."
• "Make a duplicate of the Security Reviews report for every unique value in the Department field, each filtered to its own department."

Get insights from your data. Ask it to explain a report or dashboard, aggregate data across records, combine data from multiple sources, or surface gaps, like records that are missing an expected link. For example:

• "Are there assessments not linked to a risk?"
• "How many controls failed in Q3 vs. Q4?"
• "Which open findings are linked to third parties that require periodic reassessment?"
• "What's our average time to remediate a critical finding?"
• "Which findings are owned by deactivated users?"
• "Are there any potential duplicate findings where the names sound similar?"

The agent can handle simple lookups ("How many third parties are there?") as well as more advanced requests — filtering by team or owner, grouping and ranking by a field, bucketing by time period, and multi-step questions that pull from more than one workflow.

Suggested prompts also will appear to help start the conversation if you need some recommendations. These prompts will be tailored to the particular page you are accessing the insights agent from.

Additional Examples

Permissions and safeguards

• The Insights Agent respects your existing LogicGate permissions. It can't show you a record you can't already see, and if you don't have access to any of the underlying records a question depends on, it will let you know rather than guessing. It also can't edit a report you don't already have edit access to.
• It can create and edit reports and dashboards, but it can't delete them.
• If a request would affect more than five reports at once, the agent will summarize what it's about to do and ask you to confirm before proceeding.
• Some bulk actions like bulk report creation, for example, can't be undone in bulk. If you need to undo one, you'll need to undo each affected report individually.
• The agent cannot create, edit, or delete records. It can only create or edit reports, and only if you already have permission to do so.

What it can't do yet

• Build report or visualization types that don't already exist in Risk Cloud's reporting interface.
• Send notifications outside of Risk Cloud, such as to Slack or email.
• Read data from outside Risk Cloud's own workflow data.
• Accept a file (like a PDF) in place of a written prompt.
• Share a conversation with another user, or save its text responses directly into a report or dashboard.

Giving feedback

Since the Insights Agent is still in beta, your feedback shapes what ships next. You can rate any individual response with a thumbs up or thumbs down directly in the chat and click the "Submit a Bug" to provide more feedback or to report a bug. You can also submit directly here.

Frequently asked questions

Is there any risk to my data? No. The Insights Agent can't modify or create records, only reports, and only ones you already have permission to edit or create.

What AI models does it use? OpenAI's gpt-4.1-mini and gpt-5.4.

Can I turn it off? Yes, at any time.

Does it cost anything to participate? No.

Coming in a future general availability release

Planned updates include managing Insights Agent access directly from the AI Control Center rather than through module entitlements and roles alone, and generating the existing Reporting Insights Summary through the more intelligent Insights Agent.

Our AI documentation in our trust center also applies to the Insights Agent so more information can be found there.