AI Features

Content

To setup our AI features, please see our dedicated setup guide: AI Setup

The AI feature is a secure, AI-powered assistant designed to deliver real-time, accurate answers and insights, without ever compromising data privacy, security, or compliance.​

Unlike typical AI tools, the AI feature is built with governance and visibility control at its core. It operates entirely within ​your existing Panintelligence platform and can only access data that is already visualised — in charts, dashboards, or ​reports — aligned to your Role and User Profiles, whilst embracing our Organisations and Subscription models for multi tenanted solutions. The AI feature does not query raw data sources, metadata, or database fields ​directly.​

​

Powered by a Large Language Model (LLM), the AI feature connects to your live visualisations, ​surfacing only trusted, existing data. If the data isn’t visualised, the AI feature simply returns no answer — eliminating ​the risks of AI-generated hallucinations or confabulation.​

​

Embedded seamlessly in your dashboards and reports, the AI feature provides governed ​AI-powered insights, based only on what you’ve already chosen to expose in the platform. Meaning you stay fully in control of what the AI feature can see and say.​

The examples that follow, describe how to use the Summarisation mode from both Charts and Reports.

Chart Level

Open

Once the AI features have been configured, a small Flask icon appears in the bottom right-handhandner of a Chart - as highlighted in yellow in the example above.

Upon selecting the Flask icon, the AI feature will use the LLM to produce a Summarised Response, which provides insights into the data using by the Chart shown above. The Summarised Response will make observations and make suggestions for actions you may wish to consider. In the example above, the AI feature is looking at a Scorecard Table which shows the percentage of targets met by different Product Names, as well as Policy Count and Consumer ratings.

Summarised Response

Below we see part of the output of the Summarised Response produced by the AI feature, from the Scorecard Table…

Open

… as the AI feature continues to evolve and learn, it is possible for different results and suggestions to be made over time. The text below, describes a full response that was output for the Scorecard Table

Demonstration Video

The video below briefly demonstrates the Summarisation mode in action;

Open

Demonstration - Chart Summarisation Mode

Using the AI feature in this way is a manual activity that helps end users gain insight into individual charts when needed.

Report Level

To compliment the Summarisation Mode for Charts, the same feature is available for Reports too - allowing for the Summarised Responses to be included directly within Reports.

The example below focusses on creating a Report over a Consumer Duty dashboard. For context, the Consumer Duty dashboard is designed to be used by businesses regulated by the Financial Conduct Authority (FCA), typically aligned with the Banking, Financial and Payments sector.

It is important that if you are sold a product (e.g. a loan, insurance policy, credit facility etc) that;

• The product(s) and services are appropriate to you: (you were not mis sold / over sold - for instance if you needed to make a claim - did it cover what you expected?)

• The products offer fair value (you were not taken advantage of, is the company exploiting people and charging too much?)

• Do you understand how to access the services (if you bought an insurance product and can’t make a claim when you need then it has let you down)

• How good is the support? (do you just end up in a call queue, or do the digital services help solve problems you may have?)

To ensure that businesses are fulfilling such requirements, they track data from their applications/software, take customer surveys and monitor telemetry systems to gain data that feeds into a Consumer Duty dashboard - as shown below;

Open

• Create a Report - as per the standard Report creation practices

• Including the Charts - in this example, the Balanced Scorecard chart is being used

Open

• From the right hand menu, beneath the Generate Text heading, you can see the Balanced Scorecard chart - with the Flask icon to the right

• Clicking the Flask Icon will generate the text for the chart

Open

• You can now copy and format the text to a chart cell of your choice on the report (this step will become automated in the future)

• Should you generate a PDF from the Report, this will show the Chart element, and the Summarised Response that was placed onto the Report - as shown below

Open

• When multiple Charts are required on a page/report, you can choose which chart to generate the Summairised Response for and copy this information into the fields created within your report as needed - as shown below;

Open

Assistant Mode provides the user with the ability to ask questions about existing Charts and receive suggestions along with the ability to jump to the category directly. In the example that follow, this is being performed at the overall Category level.

Category Level

Upon accessing any Category, the menu banner at the top of the category presents the icons below - with the addition of the Flask icon (highlighted in yellow).

Open

Clicking the Flask icon will engage the AI Assistant window to appear, in the bottom right hand side of the Category - as maximised below;

Open

Dashboard AI Assistant - Dialogue window

Within the text box at the bottom, users can ask the AI Chart Explorer questions about data and receive suggestions and links to relevant charts.

Demonstration Video

The video below briefly demonstrates the AI Chart Explorer mode in action.

Open

The video demonstrates a Category that contains 3 existing charts, where the user has asked the AI Chart Explorer specific questions for more Sales information. A response is provided by the AI Chart Explorer and leads to further Charts and narrative being provided for investigation and/or consideration.

The AI feature has been created with a long-term vision in mind, and as such, it is currently in its early stages of development. This feature is set to evolve over time, guided by a clearly defined roadmap of future enhancements. These improvements will be prioritised based on the feedback gathered from this initial pilot phase - from both customer and internal sources.

Customer Feedback/Requests

• Feedback and queries can be logged directly with support@panintelligence.com or through our ZenDesk portal. These will be routed to our Product & Consultancy teams for follow up, and will be handled outside of existing SLA agreements due to the nature of this being a pilot.

Before we get started it is important to be aware of the Key Considerations & Pre Requisites before you can start to use the AI features.

Data Processing & AI Responses

When you interact with the AI features and send requests to the LLM, please be aware of the following:

1. Data Transmission:

   • The data you enter will be sent to the LLM for processing - see here for more information on OpenAI models

   • This may include any text, queries, or information you provide.

2. AI-Generated Responses:

   • Responses are generated by an AI Language Model (LLM) and are based on patterns learned from vast datasets.

   • Results may not always be accurate, up-to-date, or contextually appropriate.

3. Verification Required:

   • Do not rely solely on the responses for critical decisions, legal matters, medical advice, or financial guidance.

   • Always verify information from trusted sources before taking action.

4. Data Sensitivity:

   • Avoid sharing personal, confidential, or sensitive information when interacting with the AI feature.

   • This application does not store your chat history beyond the active session.

Troubleshooting Information

• View system logs: <https://your-dashboard.example.com/pi/log> (admin access required)

• More diagnostics: <https://your-dashboard.example.com/pi/diagnostics>