One AI Assistant User Guide

About One AI Assistant

Overview

One AI Assistant is an exciting new generative AI feature in One Model for generating charts, tables, and values from your organization’s  people data by simply asking questions or using text prompts. Prompts  can be as simple as “headcount” or as complex as “show me hires and terminations for the past 12 months broken out by gender for the finance org unit and include a forecast”. It also provides an intuitive yet powerful interface for interacting with your data, such as adding metrics or dimensions and modifying selections.

One AI Assistant Capabilities

• Automatic data visualization selection
  • A bar, spline, value, or table may be automatically selected based on the shape of the data or the specific suggestion in the prompt
  • By default, a metric by a dimension generates a bar chart
  • By default, a time series is displayed as a spline chart
  • Explicitly requesting any of the following visualization types is possible (example: "headcount by location as an area spline chart"):
    • spline
    • bar
    • line
    • area
    • area spline
• Metric selection
  • Multiple metrics can be selected from a single prompt
• Time selection(s)
  • Explicit selections can be requested
  • Time models such as “Today” or “Past 12 Months” can also be requested
  • Past X months/quarters/years can be requested, even if not available as a time model
  • Last month/quarter/year can be requested, even if not available as a time model
• Dimension selections
  • In addition to the dimension, level selections can be requested
  • Explicit node selections are also possible
• Dimension filtering
  • Inclusions and/or exclusions can be specified
• Metric sorting and filtering
  • You can request things like "which 3 job families had the most terminations last year"
  • Most, least, highest, and lowest are good words to use to get these results
  • If a number is not specified, 5 is the default
  • Sorted and filtered charts automatically display as horizontal bar charts
• Drill through to detail
  • Drill through is available by selecting a value in a table or chart
• Forecasting
  • Number of periods can be specified in the prompt
  • Other forecast settings are not available in the assistant yet
• Feedback
  • Feedback provided within the assistant is used to improve the product
  • When providing negative feedback, the option to add comments is available

Using One AI Assistant

Please note that while more functionality is coming, One AI Assistant currently generates values, charts, and tables from text prompts. Requests outside this scope may result in an error message or the assistant will attempt to generate a chart that will not be what you asked for.

1. Accessing One AI Assistant and the Initial Prompt

1. One AI Assistant can be accessed from anywhere in One Model if it is enabled for your instance and Application Access role, identified by a prompt input control in the page header.
   
2. There are two options for proceeding:
   1. Enter a question or request and click Go or press Enter.
      
   2. Hover over the page header to reveal an expander. Click the down arrow to view and select example prompts, or enter your own question and click Go or press Enter.
      
   3. One AI Assistant will leverage generative AI to identify metrics, dimensions, dimension levels, time selections, filters, sorting, and forecasting in your prompt and translate those into a data visualization. Examples of prompt patterns can be found here. For practical tips and things to consider while creating prompts, check out our One AI Assistant Best Practices article.

2. Interacting With the Data

Just like Storyboards in One Model, the data displayed in One AI Assistant is interactive. 

• In basic charts, hovering over a data point displays additional information about that data point. Clicking a series in the legend hides that series in the chart.
• The ability to switch to a table view is always available.
• Drillthrough to detail is available as it would be in a Storyboard. Clicking on a data point displays the detailed data behind that number. What is displayed here depends on the way you have One Model drill through configured.

3. Providing Feedback 

Providing feedback before modifying your prompt or query is encouraged. Simply select a thumbs up or thumbs down below the chart to let One AI Assistant know if it got the result right or wrong. If a thumbs down is selected, you can submit additional comments that will be used to improve the performance of the assistant.

Feedback can only be submitted once per prompt, but editing or typing and resubmitting allows for new feedback. Your feedback is reviewed by the One Model team to enhance the assistant’s performance.

4. Modifying Your Prompt

Modifying your prompt is as easy as editing or replacing the text above the data visualization and clicking Go or Enter.

When you submit a new prompt the feedback controls are reset and you are able to provide feedback again. Thumbs up indicate a correct result while thumbs down means the assistant got it wrong.

5. Modifying the Query

One AI Assistant features robust and intuitive query definition capabilities. These capabilities are explained below. After making any change to the query, the Run Query button will be enabled. Clicking this button applies the change(s) and re-runs the query.

1. Data Visualization - Select from Basic Chart, List Table, or Key Value. Key Value can only be generated for queries that display a single value. Please note that several chart types fall under the Basic Chart option. These can be defined for each metric as explained in the next bullet.
2. Metrics 
   1. Metrics can be added using the + button or removed using the x button on each metric chip. Multiple metrics can be displayed on a single chart or table.
   2. When Basic Chart is selected as the Data Visualization, clicking on a metric chip for a metric already included displays the following options:
      1. Series Type - Series types include Spline, Column, Line, Area, and Area Spline. A single chart can contain metrics that have different Series Type selections applied. Bar/Line is an example.
      2. Horizontal Bars - By default bars in Column charts are vertically oriented. Selecting this option changes them to horizontal.
      3. Filtering - In basic charts where a single metric is included, the results can be filtered by Top or Bottom
         1. Count - If Filtering is enabled, a count of records to return can be defined.
      4. Sorting - In basic charts where a single metric is included, the results can be sorted ascending or descending.
3. Dimensions - Dimensions display on the Y axis of charts and rows in tables.
   1. Dimensions can be added using the + button or removed using the x button on each dimension chip. Multiple dimensions can be displayed on a single chart or table.
   2. Clicking on a dimension chip for a dimension already included displays the following options:
      1. Pivot and show label on X-axis - Toggling this on results in the dimension being moved to Pivoted Dimensions.
      2. Selections - The selected selected, included, or excluded dimension level(s) and/or node(s) are displayed here, along with an option to Clear All selections.
      3. Levels - Clicking on levels includes or excludes them from the visualization.
      4. Nodes
         1. The eyeball icon indicates selection of a node.
         2. The checkmark icon indicates inclusion of a node.
         3. The X icon indicated exclusion of a node.
         4. Clicking the right arrow if one is displayed expands that node of the dimension to display nodes included at the next level.
      5. Time Models - Exclusive to time dimensions, Time Models enable dynamic time selections that automatically update as time progresses. If Today is selected, every time the site is re-processed, Today will be updated to the current date.
4. Pivoted Dimensions - Pivoted dimensions display on the X axis of charts and columns in tables. All other settings are the same as regular dimensions.
5. Pivoted Columns - Pivoted columns allow you to include additional columns from the fact and dimension tables from the data model in charts and tables in One AI Assistant. They render the same as pivoted dimensions do but come from raw columns rather than structured dimensions.

6. Exporting and Pinning Results

Saving and sharing information is possible directly from One AI Assistant.

1. Export Chart generates and downloads a PNG image of the chart displayed in One AI Assistant. Please note that this is only an option for Basic Chart types.
2. Pin to… allows for the following options:
   1. Pin the chart or table to the Insight Library where it can be accessed from Explore.
   2. Create a new Storyboard and pin the chart or table to it as a tile. If this option is selected, you will be prompted to name the Storyboard, the page, and the tile.
   3. Pin the chart or table to an existing Storyboard as a tile. If this option is selected, you will be prompted to select a page and name the tile.

Examples of Prompt Patterns

Substitute your metric names, dimension names, and time selections for the bracketed text in these examples. These can be combined into more complex prompts.

• [metric name(s)]
• [metric name] and [metric name]
• [metric name(s)] for [time selection(s)]
• [metric name(s)] by/for [time trend]
• [metric name(s)] by [dimension name(s)]
• [metric name] for [dimension selection(s)]
• top/bottom/highest/lowest [dimension name] for [metric name]
• top/bottom/highest/lowest [number] [dimension name] for [metric name]
• forecast [metric name(s)]
• [metric name(s)] by/for [time trend] and include a forecast

Prompting Best Practices

Since the assistant relies on a large language model (LLM) and a vector database to interpret prompts and match them with your data, using clear, deliberate prompts helps ensure accurate results. Here are some best practices to help you accomplish this.

1. Be specific, yet concise: Clearly state the metrics, dimensions, and time ranges you want to analyze. Avoid vague requests like "show performance" and instead ask for "show headcount for the last 12 months by department." While it's important to be specific, avoid overloading your prompt with unnecessary details. A clear, concise prompt is easier for the LLM to interpret correctly.
2. Dimension levels: If you know the specific dimension level you want to see the metric by, include it in your prompt (e.g., “show terminations for the last 5 years by sup org L3”). Otherwise, the dimension will default to level 1, and you’ll need to modify the query in the query panel to select the appropriate level. 
3. Use familiar terms: While the vector database can understand different ways of referring to common metrics and dimensions, using terms that are exact or as close as possible to those in your data will improve accuracy (e.g., if your data uses "location," prompt with that instead of "office site").
4. Avoid ambiguity: If there are similarly named metrics and dimensions (e.g., "work location" and "home location"), make sure to specify which one you mean. Avoid terms that could apply to multiple fields when possible.  
   • • The One AI team is also developing a configuration feature for site administrators to prioritize specific metrics and dimensions. For example, if your instance has multiple Headcount (EOP) metrics (e.g., Headcount (EOP), Headcount (EOP) - Women, Headcount (EOP) - Managers), admins can prioritize which version of "Headcount" is returned when users simply prompt "headcount."

5. Include context: When needed, add additional information to narrow down the results. For example, specify time periods ("last quarter" or "past year") and organizational units (e.g., "finance team" or "sales department") for more precise outputs.
6. Refine as needed: If the initial output isn’t exactly what you want, rephrase the prompt or add more details. For instance, if the result is too broad, narrow it down by including additional filters. If you have drill through access,  drill through data points in the query view the metric definition to validate when you are unsure if the assistant selected the correct metric.

Next Steps

Unlocking the full potential of One AI Assistant allows you to turn data into clear, actionable insights that drive impactful decisions. To continue your learning, explore our additional resources:

• How One AI Assistant Works: Understand how the assistant processes prompts and leverages generative AI, taking you from a question or request to a complete data visualization.
• One AI Assistant Configuration Guide: Designed for admins, this guide covers initial setup and adjustments, including permissions, example prompts, and vector configurations.