In the current release, we’re introducing the first version of AI-powered template creation and filling.

The goal is simple: help users go from an idea and a sample PDF to a working, data-driven template in minutes, without needing CSS skills or deep Salesforce knowledge.

AI as your new template author
We’re introducing a new way to create templates where AI does the heavy lifting, and you fine-tune the result.

New template creation flows
You’ll see a redesigned “Create New Template” experience that lets you decide what you need first: printable materials, email layout, or landing page.

The next step is to select the way you want to get the template:

• Create from PDF (AI) - available for Printable templates only.
  Upload a sample PDF (e.g., a brochure or report), and AI will:

            ◦ Reproduce the layout in HTML

            ◦ Detect sections, including multi-page layouts

            ◦ Identify text, images, and maps as separate elements

            ◦ Build a data schema for these elements 
              This becomes the AI Template that you can edit and connect to Salesforce.

• Build from Scratch
  Start from a blank printable template to create the template w/o a prototype.
• Start from Library
  Use the prebuilt default templates as a starting point and adapt them to your needs.

What is needed to create from PDF

Upload a PDF
To create a printable template with AI, upload a PDF sample on the 1st step.

There is a dedicated Data Security Policy for the AI Template that explains what is sent to AI models, how it’s stored, and the safeguards we use. Read it via the link.

Select the Main Object
Select the main object from the dropdown of all Salesforce objects in the organization, choose from the frequently used ones, or assign this task to AI.

The system now analyzes the PDF content, understands the context, identifies the most suitable Main Objects from your organization (e.g., Property, Account, Listing), and suggests the TOP-3 most probable options, including the confidence level and the reasoning behind why each Salesforce Object fits. Optionally describe what your PDF is about to help AI understand the context better.

Smart variable detection & mapping to Salesforce

Variables states
The elements in Ascendix Composer have 3 states:

• Static element: the element that looks the same in every generated report. It is set manually and once. If AI misdetermined an element as static, but it should be dynamic, click the “Link field button” in the Edit tab and assign it a path to Salesforce via natural language instructions.

• Dynamic unmapped variable: a dynamic content w/o set path for data grabbing in Salesforce. There are 2 options:

         ◦ keep it static if AI misdetermines the field, but it does not imply the dynamic content;

         ◦ build a path for data grabbing, the same as for static elements initially;

• Dynamic mapped variable: a dynamic content that implies pulling data from Salesforce and can be changed from record to record. If this element is misconfigured and should not be dynamic, keep it static in the Edit tab. In this case, the data from the source PDF will be substituted.

Text variables

• AI scans your uploaded PDF / generated layout and detects dynamic text fields that are likely to be driven by Salesforce.

Source PDF
 

Resulting template

• These variables are surfaced in a dedicated Edit tab in the AI Template editor, where you can:

           ◦  Review how AI mapped each variable to a Salesforce field

           ◦  Adjust mappings addressing the request to AI where needed

           ◦  See confidence levels and handle edge cases where AI is not sure

Images

• AI now recognizes images in PDFs and classifies them as:

         ◦  Static images (e.g., logos, icons, key feature images)

         ◦  Dynamic images that should be driven by Salesforce files or attachments

• In the editor, you get a dedicated UI for image variables where you can:

          ◦  Link dynamic images to Salesforce Files or Notes & Attachments

          ◦  Manage fallback images

          ◦  Adjust how images fit within the layout

Maps
Real-estate-centric templates often rely on maps, so we’ve taught AI to understand them, too.

• AI detects maps in your PDF and converts them into Composer Map elements rather than generic images.

• A dedicated Map Edit UI lets you:

         ◦  Bind the map to a Salesforce location field

         ◦  Control how maps appear in the final PDF

         ◦  See and preserve map bindings across sessions.

From sample to data-driven publication
Once the AI Template is created and mapped, you can generate actual publications from Salesforce data.

Automatic filling with real data
AI-powered templates can now be filled with live Salesforce data, replacing dynamic variables and executing loop patterns for collections (e.g., lists of properties, etc.). The operation goes on in the background, so you can leave the platform or interact with another template case.

Publication generation & export

• From the Composer UI, you can generate publications from AI Templates just like from classic templates.
• Publications generated from AI Templates can be exported to PDF, preserving maps, images, and layout as much as possible within PDF rendering capabilities.