AI and Agentic Automation

Hi Community, I am encountering a strange issue with Record Chat. Occasionally, I receive unintended or unusual responses. The behavior is inconsistent—sometimes it occurs after a few messages, other times after just one or two, and occasionally even with the very first message. I’m unsure whether this is a bug or a configuration issue. One example, shown in the attached image, is that the conversation output includes entries formatted as 'H:' and 'A:', resembling a question and answer dialogue, which I did not configure or expect. My setup appears straightforward: I’m using fields from both the primary record and the required related record fields, passing the correct identifier. I also attempted to adjust the behavior using the additionalInstructions parameter, but this did not resolve the issue. Has anyone else experienced this problem or have any insights into what might be causing it? Thank you in advance!

Hi all,  Is there any way that we can implement a copilot that will answer about our record data? For ex: it should be like a chat bot and it should be in the record summary page, if we ask about the details of the record summary, it should provide the answer. Can anyone please help me in implementing this ?Thanks in advance!!

Hi, I'm trying to Extract data from PDF and store those data in Database. I gave 8 fields in the extraction. I'm able to see all my fields in the ai skill extraction but in the Database I didn't get EMPID, EMP type. Anyone please tell me the reason why I'm not getting those values in my Database. Thank you in advance.

Hi All, Has anyone successfully deployed AI Document Center 2.9 on Appian 25.2?https://docs.appian.com/suite/help/25.2/aidc-2.9/whats-new.html When I inspect the new package: AiDocumentCenterv2.9.0.zip, I receive the following error message:The processModel [id= uuid=0005ee34-9c41-8000-0422-7f0000014e7a] was not imported because a required precedent is missing: The referenced smart service [local-id=internal3.rs2_ai_skill_advanced_idp_tools3, name=Advanced IDP Tools] was not found. (APNX-1-4073-000) (APNX-1-4071-006) When I add the 'Advanced IDP Tools' smart service to a new process model and check what version is created, I can see that I am running the following version of the smart service:Advanced IDP Tools (internal3.rs2_ai_skill_advanced_idp_tools2) It looks like my Appian instance is running an older version of the Advanced IDP Tools smart service, does anyone know how I can update the Advanced IDP Tools smart service to the latest version? Kim.

"RUN EXTRACTION" button is not visible on AI Document Center, after creating an Extraction Model. Why?

"RUN EXTRACTION" button is not visible on AI Document Center, after creating an Extraction Model. Why?

Hi everyone, good day. I’m working with an AI Skill for document extraction in Appian, specifically targeting a structured and standardized document used in Mexico: the Constancia de Situación Fiscal (a tax ID certificate). My goal is to extract the following fields: RFC (tax ID) Business or legal name (Razón Social) Start date of activities List of economic activities (shown as a table in the document) The issue I’m facing is that the AI Skill successfully extracts all the individual fields (RFC, name, start date), but it completely skips the table that contains the economic activities. I’ve already configured the document type, field mappings, and extraction zones where possible — I’ll share screenshots of my current setup below. Has anyone encountered a similar issue with table extraction in structured documents? Any suggestions would be greatly appreciated. Best regards, Víctor Valenzuela

Hi All, I am using the Advanced IDP to extract Key-Value pairs from a PDF. Even though the key - values are present in a 2 columns table, I decided to use Key-Value extraction instead of table extraction as each key occurs only once. I can see the extracted values properly in the IDP object, in the output of IDP in the process, but whenever I am assigning the values to individual variables or CDT inside a process, lot of additional semi colons are getting appended along with repetition of the main value. I could see a pattern - if the key is available in the 3rd row of the table, then 2 semi colons are getting appended before the value, if the key is in 4th row, then 3 semi colons and so on. Example: ; ; abcde; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; ; abcde; ; ; ; Any idea?

Our team is exploring the use of the a!recordsChatField and the new smart search capability.  Our use case: Business is using a web page with lots of links to access apps (both Appian and non-Appian). Users would like to be able to find the app they need without having to scroll through the "link farm". We have built a form in Appian to provide a better user experience and would like to add an AI component that will allow users to find the link they need quickly.  We have tried using smart search but are finding it is somewhat limited and does not provide the results we are looking for, plus the text output is rather lengthy. We are looking for a one-sentence result instead of multiple paragraphs of information. We have a record that contains the list of all the apps, their descriptions, and their urls. a!recordsChatField provides much better results by providing the name of the app, its description and its url, however it would be great if in the results

I'm getting an error when trying to publish a model using AI Skills (Extraction). Could this be a restriction of the Community Edition, or is there something I need to configure?

I am using AIKnowledgeAssistantField() smart service And here I want the history of the queries asked and the response got from the AI AssestantWe have a field to capture the history i.e "conversationValue"But how to we use if is not properly mentioned anywhere My requirement it to upload the document and ask few queries related to the document and then capture the Audit history so i want all the queries and the responses saved somewhere.I tried using Document Chatbot but there also i did not find a way to save the history or there are any logs that have been captured. If its not possible in AIKnowledgeAssistantField is there any other way we can get the Audit history 

Is there way in prompt builder to pass it a document to reference? example: Based on the policies in Document_XZY does the data being passed in in Json format appear to be against any of the documented policies?

I have a list like {1,2,3,4}, I want to add new items like {1,2,3,4,5,6,.....} in the same list without using a new variable to store the new list.

Hello team, We have a requirement to implement an chatbot in Appian and similar to Enterprise copilot. Will it be possible to use this Appian LLM standalone whenever needed? So that we can make the bot LLM configurable. We have to give option to connect with Appian LLM / OpenAI / Azure OpenAI. Thanks in Advance!

Hi team, where we can find the sample invoice documents in our Appian documentation? to train the document classification model. thanks in advance!

I have a structure of record type A like this: {id: other business field:, relationships: record Type2: { id, A_id, other business field } recordType3:  { id, A_id, other business field } }, After user change the data, I want to check whether the business field is actually modified. What I thought is to traverse the list of this record and construct new objects using the business field, use the new objects to compare. I want to know is this the only way to solve my problem? Is there any easy way can solve this.  

I'm encountering an issue where I need my suggestion function to return both a dataSubset and an output value (like a flag). The goal is to allow users to enter a new project code if it doesn't exist and make dependent fields editable. I need a way to pass that output from the rule back to the interface Trying with a!update

In a queryentity function, in an interface, I'm getting an error like "Unexpected error executing query (type: [CPVAccountsStatementTrackerDT8331], query:[queryentity expression], order by: [[Sort[FinancialInstitutionOther asc]]], filters:[(firmId = TypedValue[it=1,v=-1)])". How should I resolve this error. In one of filters, we are filtering based on -1 but this is working in some environment while failing at  some.

Hi, I have a RPA that takes files from my local machine and upload them into an Appian folder with the Upload Document module provided in Appian Services.. When I run the RPA within the Appian RPA console it works perfectly fine. I want to trigger this RPA from an Appian Process Model that uses the Robotic Process Smart Service but when I do so the RPA start fine but then I have an error that I can't solve so far... (looks like it's about the Target Folder).Error code APNX-1-4530-021 with message "docFolderUuid query paramter is not found." found for resource /api/v1/uploadDoc Can anyone help me? Thanks, Jeremy.

I'd like to understand how the count of descendant rules affects interface performance and how this count is calculated.

HiCan we extract data from unstructured documents using Appian IDP Gen AI features

Hi, I'm looking for more details about IDP, implementation directions and where can I find more help/documentation regarding this. I am new to IDP anything that you can provide would be helpful.  Thank you.

Hi everyone,Is there any way to maintain an audit trail of user queries and responses monitoring for review purposes of any chat bot mentioned above.thanks in advanced

Hi AllWhen i am doing filtering not getting expected output.Below code is mentioneda!queryEntity( entity: cons!LOCALE_TRANSLATION_NEW, query: a!query( selection: if( rule!P3_utils_checkIsNullOrEmpty(ri!columns), {}, a!querySelection( columns: a!forEach( items: ri!columns, expression: a!queryColumn(field: fv!item, alias: fv!item) ) ) ), logicalExpression: a!queryLogicalExpression( operator: "AND", filters: { a!queryFilter( field: "label_txt", operator: "=", value: ri!label_txt, ), }, ignoreFiltersWithEmptyValues: true ), pagingInfo: if( rule!P3_utils_checkIsNullOrEmpty(ri!pagingInfo), a!pagingInfo( startIndex: 1, batchSize: cons!P3_BATCH_SIZE[7] ), ri!pagingInfo ) ), fetchTotalCount: and(ri!isFetchTotalCount))=======================================When i searched "Research". It is giving me expected outputIs it default behavior of Appian ? Expected result - It should not return anything because there is no label like "Research  "

Hi everyone, I have a challenge where I need to classify more than 10 different types of documents. I understand that the traditional Classify Document approach in Appian requires a substantial amount of training data to build an accurate model. Is it possible to use generative AI (Gen-AI) for document classification in Appian instead? Or is Gen-AI generally not suitable for handling multiple document types like this? I’m aware that the current Appian AI Skills for document classification do not use Gen-AI, but I wanted to check if there are any alternative approaches or best practices. I also tried using extraction with unstructured Gen-AI for classification, but it turned out to be more complex than I expected—especially converting the output to JSON and so on. Thanks for any advice or suggestions!ps; currently lacks of documents so i not using the classification document method xD