AI and Agentic Automation

Hi all, I am using the AI Doc center version 2.9. I am trying to extract invoices with single and multiple line items. There is a field called service code, which needs to be a list of string. In the extraction details, I have given an options rule, which returns a list of string values, based on the type of invoice.  The prompt for this field is: " This should be selected from the values in the options. This value should be null if it is a dental receipt.  Please ensure you return a text value. When it is a single value, please ensure you return a single text value." The extraction is working fine when there are multiple line items in the receipt. If it is a single line item, the value is extracted as {"[value:Chiro-Init", "spanId:99492f14]"}. I need the value to be "Chiro-Init". Due to this, the extraction summary page is displaying the error: "A rich text item or rich text list item in a rich text display component [l

Hi Everyone, I was trying to explore Appian integration with Gemini to summarize document, email and chatbot interface. If anyone did this either with or without plugin,  let me know the details in brief how you achieved it. Also, did you faced any limitations? (PS: Specifically needed to integrate with Gemini)

Getting this error using a!documentsChatField when a basic user is accessing it on the interface. Users have access to relevant groups but not able to access the document chatbot. Is there any AI specific group that the user should have access to?

If I run extraction again and again for same document, still confidence score is same for that field. If confidence score is greater than threshold, want to skip reconciliation and map the details. How will confidence score improve for a field in document extraction in docCenter? 

Hi All, I'm trying to learn Document Extraction using AI Skill and I'm facing an issue described below: Step 1: I have created an AI Skill object for Document Extraction and I have added only two fields in it. (Screenshot below). Step 2 : I have created a model to add this AI Skill(Ignored the XOR checks before reconciliation as it's just for learning purpose) Issue: The node stays active for a long time and I get an error "The number of retries to complete the execution of the smart service activity has been exceeded." The document is pdf and it has not exceeded the size limit. What is going wrong here?

I am working on Excel classification using AI Document Center .After the classification process completes, I m not getting the confidence score in Pv  and  database. how to resolve it?

Do we require this AI Document Center - Generated Objects application to add the model to the Appian Process?

I would like to implement a function in Appian that retrieves a list of actions associated with a record. I have been testing multiple approaches and configurations in an attempt to achieve this functionality; however, none of them have produced the expected results so far. Despite extensive testing and exploration of different alternatives, I have not yet been able to successfully obtain the desired list of record actions.

Hi Everyone, Thanks in advance, I have list which is having list of dates {1/1/2026; 2/2/2026; 12/31/2025; 11/24/2025}.I needs to identify the least date and Maximum date from the list. Regards Divya K

Hi everyone, I’m currently experiencing an issue where the round() function does not return the expected results. Below are some examples. In case 2, the expected result is 512.93, but Appian returns 512.92. Additionally, in case 3, even when applying the workaround recommended in KB-1426, the output is still not correct. KB1426 document : community.appian.com/.../kb-1426-unexpected-results-on-decimal-calculations-and-rounding If the round() function cannot reliably produce correct rounding results in cases like these, is there any known alternative approach — such as a custom Expression Rule or reusable utility — to achieve accurate decimal rounding in Appian? I’m looking for: A method that consistently returns mathematically correct rounding results A reusable rule that avoids floating-point precision issues Any best practices other users may have implemented for similar cases Any suggestions, patterns, or examples would be greatly appreciated. Thank

I have decision rule that's need to have output list of text and have more than 500 characters. I am using the output of desicion rule into multiple dropdown and store values into database that have limitation 1000 characters. How to increase this limitation?

Hello, I'm exploring AI Agent features in Appian 25.4 release and going through this tutorial https://docs.appian.com/suite/help/25.4/build-triage-agent-tutorial.html and trying to recreate this. For this, I tried calling the expression rule and save the result in output. While testing, it just shows processing for long time. Kindly help me with what I'm missing here. Is there is any specific prompt do we need to follow to map the inputs or the AI agent will automatically infer the data based on our prompt.

Hi Appian Community, i watched the 25.4 release showcase by product manager for a!chatField component and custom chat interface  from timestamp 54:37 and wondering how to achieve the same functionality which i can further modify as per my use case. i tried using custom chat interface present in appian documentation but it is limiting me to only use a card interface once and it will show same in every response.My goals:- Achieve same functionality of booking an appointment via chat field or custom chat interface. it should be intelligent to dynamically respond as per user, such that when first given symptoms it only shown a text response, but later responses after the appointment confirmation it showed user a custom card layout with a confirm button, and later another custom card layout with confirmation and another user input text field where user can enter further symptoms and upload any images(optional).  i have uploaded images in sequence from video, in which the respon

As we are having a Advance License of Appian but we are not getting option to how to Enable this in our environment is thing need to be done via raising a support case in myappian or it should be done via Admin Console. happy to hear any suggestion for same thanks 

Hi Appian Community,I am using AI Document Center in Appian to Extract an Excel file. The extraction runs successfully, but the Extracted Value column is empty. Only the field labels are displayed (e.g., S.No, Name, Age, Location, Role, Salary, Time, Status), but no actual data is populated and also not able see the preview of excel  Any guidance or best practices would be appreciated.

i have a sentence that "Benefit Period: To Age 75; Indexation: CPI Linked" in this I need to validate that whether "Indexation: CPI Linked" is found or not  please give me a suggestion

Requirement: User wants to search across the PDF documents. Currently we store PDF documents generated via HTML in Appian. User wants to have a search box and search through all these PDF documents and return all the documents having matching text. Just wanted to know if there are any OOB features that supports this requirement.  Any suggestions would be appreciated. Thanks in advance.

Hi everyone, I'm having an issue with a grid in my interface. When using System Administrator account, the data loads correctly when the next page button is clicked.  But when using a Basic User account, the grid shows the same first-page data, and the paging control is not updating as well. I checked all related objects used in the code, all have Viewer access for the Basic User's group. Code snippet: a!gridField_23r3( data: local!grid.data, pageSize: 5, pagingSaveInto: fv!pagingInfo, validations: {} ) Thanks in advance for any insights!

Hi All, In database table, i have one field called 'versions' which will be captured from agent run. So when the agent runs, it will give some array of outputs for each version like below, 1st time run Output {workId - 1, decision - need referal, status - pass} {workId - 2, decision - no referral, status - pass} {workId - 3 , decision - need referral, status - pass} 2nd time run Output: {workId - 1, decision - no referal, status - pass} {workId - 2, decision - need referral, status - pass} {workId - 3 , decision - need referral, status - pass} {workId - 4 , decision - need referral, status - pass} So, now I have to compare workIds result between 1st run output and 2nd run output and have to show below 3 sections, Section 1: New Referral - which will have the workIds which needs referral in output 2 but not in output 1, so from above example it should be below rows from output2, {workId - 2, decision - need referral, status - pass} {workId - 4 , decision - n

Hello, again. This is a strange one.... and I'm on version 24.1.281.0 I have two expression rules... one generates some error messages (or none) and the other deals with the errors.  However, in the scenario where the first expression rule generates no errors, the second one still thinks there are errors. Expression 1: (called SH_ErrorGenerator) a!localVariables( local!stringArray: {"a", "b"}, local!error1: "", /* value from another expression rule normally */ local!error2: "", /* value from another expression rule normally */ local!error3: {""}, /* value from another expression rule normally */ local!errors: { local!error1, local!error2, local!error3 }, local!allErrors: cast( /* make sure this is list of string not list of variant */ typeof(local!stringArray), reject(a!isNullOrEmpty, a!flatten(local!errors)) ), local!anyErrors: and(not(isnull(local!allErrors)), length(local!allErrors) > 0), { errors: local!allErrors,

Hi everyone, I am encountering the following error while accessing a Record Type in one of my applications: “The configured database table is invalid. The XXXX_RF_EMP_DETAIL_STATUS table does not exist in the configured data source. Change the database table to continue.” Could anyone please guide me on how I can retrieve the missing database or what best practices I should follow in the future to ensure the database is properly saved and linked? Thank you in advance for your help!

Hi, I am new to using the interface component a!dataFabricChatField.  ( I've used record chat component, and that worked as expected. ) With Data Fabric chatbot, how should we implement security?  We only want the users using the chatbot to have access to recordType records that pertain to their company.  I tried using the instructions field in the chatbot component to specify which company the logged in user belongs to, but that did not work either. What is/are the best to accomplish that? Thanks.

In the development environment I have installed two plugins Google Maps Connected System and GoogleMaps (the latter gives me the map and direction fields components). When making the transition to the UAT environment, only the Google Maps Connected System was deployed correctly, however the second plugin was not, I checked the logs and did not see any errors, could you help me. THIS IMAGE IS RELATED TO THE DEPLOYMENT OF THE GOOGLEMAPS CONNECTED SYSTEM PLUGIN THIS IMAGE IS RELATED TO THE DEPLOYMENT OF THE GOOGLEMAPS FIELDS

I have a requirement where similar nested if conditions are required in multiple interfaces. So I'm trying to build a decision rule so that if there is any update required it can be done in the rule and it can reflect in all interfaces. But the outputs are list of ids which are stored across different constants for each condition. For example, If condition 1 satisfies I want the output to be constant1 and if condition 2 satisfies I want the output to be consant2. Each constant has more than 50 ids and I want the output to be in the same order they are present in the constant. While configuring the decision rule I'm able to see all the ids from all the constants and I have to manually select them for each condition and that too in the same order I want the output to be in. Is there any easier way or is there any way to give the constant directly as output? For now I have created expression rule and used a!match to make it simple but curious to know the above mentioned scenario.

I am trying to use AI Skill for Document extraction , I have 10 different pdf document having key  : value pairs different for eg one document has Name : XXXXX another one has   Employee : XXXXX , another one document with no name , it just say Patient name : Can we use them to train the model , or we need to have same Key value pairs When extracting the data ?  we should classify those as different documents and then process separately to extract , please provide your input ....