Use ElementsGPT to automatically generate a UPN (Universal Process Notation) diagram, allowing you to document your business processes in a clear, easy to understand format. ElementsGPT leverages AI to further accelerate your business analysis by generating maps in a matter of minutes, giving you a solid foundation for process analysis.

To generate a new diagram, go to the "Diagrams" tab in your Space.

From here, select the "New Map" button, and from the dropdown menu, select "Generate process with AI".

You can also switch between AI generation for a blank process diagram by toggling the "Auto-generate" toggle on. When active, you will see new fields displayed in the window.

For best results, fill in the map name, description, and provide further instructions.

The "Description" and "Instructions" fields are not mandatory, but it is strongly suggested you fill in as many details as possible about the map. Content that you enter in the "Description" field is saved as your map description. Content in the "Instructions" field is not saved after being used for the prompt.

See our tips below for further guidance on getting the most out of this feature.

Once filled in, select the "Create" button, sit back, and relax!

The process should take around 1 to 2 minutes depending on the length of context/size of diagram being generated.

You may need to rearrange some boxes and arrows once generated, but the idea is to give you a head-start in business analysis and process mapping using UPN.

You can also upload up to 4 images as additional context for your diagram. This could include existing diagrams in other formats (BPMN, Flowchart, etc.), written notes, hand drawn process brainstorming, and more. Supported file formats include JPEG and PNG.

Simply click "Upload File" or drag the image to the grey box in the window.

We recommend adding in the "Instructions" field some information about the provided images, and how you want them to be used. For example, if you want to merge multiple processes, or use a process as guidance to create a new one, specify that in the text you provide. It is also valuable to describe more about the image, such as the process type, and notation information, if relevant.

Example: "You have been provided with a flowchart that shows the process for logging in to an application. Use the key activities and outcomes shown in the flowchart to generate a process. Ignore [X], but include [Y] in the final process."
​

To automatically generate a drilldown on an existing diagram, right-click on the activity that you want to create a drilldown for. From the opened context menu, select "Auto-generate" and this will show a new window for entering the drilldown name, description, and instructions.

These fields are the same as the ones described in How to Generate a Diagram (above). Once filled in, press "Create" and a drilldown diagram will be generated at that level according to the provided information.

Resources are usually generated with the diagram, defaulting to "Human" resource types. In most cases, a new resource will be automatically created and added to your Resource list. If the resource already exists it will be used, rather than creating a duplicate.

You can choose for the generated diagram to only contain resources that already exist in your space. Turn on the "Use only existing resources" toggle to activate this mode.

Your generated diagram will then only add resources if they exist already in your library. Any new resources won't be added, and so you will notice some activities won't have resources.

In the event that a diagram fails to generate you have the option to retry generation. Click "Retry" on the notification and you will be redirected to the starting window again, with your previous instructions pre-filled.

We recommend amending the instructions to see if this yields better results. Tip: Lots of bullet points can sometimes be interpreted as an activity per bullet point, and too many activities can cause errors.

The instructions field has got a 20,000 character limit for both new diagrams and new drilldowns.

A maximum of 4 images can be uploaded as context.

If you are providing your own API token in Elements, it must be a compatible model token for using vision. See the OpenAI guidance for more information on which model to choose.