AI and Agentic Automation

I added a chat agent to our case summary page with a!agentChatField. When I test it as a developer it works fine, pulls case data and starts the update process. When a business user opens the same page and asks the same question, the agent fails right away with an access error. The user can open the record, so record security looks fine.

I am trying to set up the Appian Dev MCP plugin (v26.9.10) to connect an external AI assistant. The plugin is successfully installed and active in our cloud environment, and the basic OAuth configurations are already in place.However, when the AI assistant tries to connect, the endpoints /suite/plugins/servlet/stateless/lcp-api/* (for example, /lcp-api/meta) return an HTTP 404 error.Upon inspecting the application log at the /stateless/lcp-debug/ page, I noticed this flag is set to false:ae.dev-copilot-api.lcp-api-plugin = falseThe main "Appian MCP Server" toggle is already enabled in the Admin Console.Has anyone else encountered this issue? I would like to know if this specific flag implies there is a backend feature toggle that Appian Support needs to activate for our site, or if there is another configuration step we might be missing to resolve the 404 error.Any guidance would be highly appreciated!

Hi Appian Community,I am working on an AI-powered document extraction process in Appian 26.9. I have created an AI Skill to extract information from different types of insurance-related documents, such as: Police Reports Driving Licences Vehicle Registration Certificates (RC) Repair Invoices The AI Skill is successfully extracting the required information, but I am facing an issue with the output format.1. Issue with the extracted outputInstead of returning properly structured data, the AI Skill is returning all the extracted field names and their corresponding values as a single concatenated text string.For example, for a Driving Licence document, I am getting the following output:LicenseNumberVehicleTypeVehicleRegistrationNumberRcNumberDateNameDL-TRN-RJ-0002-263Scooter3RJ-27-TR-90023RC-TRN-RJ-0002-26314 October 20263Aisha Khan3As you can see, the field names and values are joined together without any separators or proper formatting.2. Unexpected number '3' in the outputI have a

For greenfield, Composer shows a Plan with all objects and how they map. Has anyone successfully done the same for an existing brownfield app and got a proper mapped plan view out of it? If yes, what approach worked?

Title: How to Extract Text from a Scanned PDF in Appian?Hi everyone,I’m trying to extract data from a scanned PDF document in Appian. Since the PDF contains scanned images rather than selectable text, I need an OCR solution to read the content and extract the required information.I tried using the Document Extraction AI Skill / Extract from Document smart service, but it is not working in my current environment.Could someone please suggest:Which smart service or plug-in is best for extracting text from scanned PDF documents in Appian? Are there any additional configurations, plug-ins, or AI capabilities required? Is there an alternative approach if Document Extraction AI Skill is not available or supported in my environment? Any suggestions or examples of how to implement this would be greatly appreciated.Thanks in advance! 

Currently when chat agent executes, in the response it shows the records queried as well. How to avoid showing records being exposed.

I wanted to share some feedback on Appian Composer after reviewing an app that was generated with it. My manager used Composer to build out an app idea, and when I went in to verify functionality, I ran into several issues worth flagging.1. Unnecessary API for existing dataComposer built an API to pull data that was already available through JDBC. This added complexity with no real benefit.2. Broken PDF generation subprocessThe subprocess responsible for generating a PDF was using an incorrectly formatted template PDF, so document generation didn’t work as expected.3. Misunderstanding of subprocess to parent data flowComposer seemed to have an incorrect understanding of how data is passed from a subprocess back to the parent process, which caused the process model to keep failing. When I instructed it to stop passing the data back from the subprocess and instead add a script task in the parent to fetch the document ID, it was not able to implement that change.4. Activity chaining remov

Hello everyone.Is there any add-on available for using voice recognition within an Appian-based interface?Thanks in advance

Creating Automation Agents

Hello, how to build an expression rule  which has 2 inputs - 1 is string, 2nd is alphabet. print the count of characters matching alphabet with the string (Umbrella to expected Result -u1m1b1r1e1l2a1)

In my development work, I’m using several tools: `dev mcp`, `mcp doc`, `dev mcp skill`, and `aurora mcp`.I’m wondering about the difference between using `dev mcp skill` and `aurora mcp`—do their functionalities overlap, or are they meant for different purposes? As I understand it, `dev mcp skill` already covers the same ground as `aurora mcp`. Could someone please clarify? https://github.com/appian-design/aurora-mcphttps://github.com/appian/dev-mcp-skills

For Appian AI Skill for Document Classification how to check, how many AI Actions it consumed. I can able to check for the GenAI skills on object itself or in output tab of the process node only but not for the classification type AI skills. Cause we have 200K Ai actions per month to spend for Appian Standard Tier. Just need to check how we can find out the expenditure of AI actions for the Classification AI Skills.

Hi ,We recently upgraded our platform to version 26.9. After the upgrade, we noticed that the AI Skill object is still showing input and output token consumption.As per our understanding, in version 26.9 it should show AI Actions consumed instead of input/output tokens.Has anyone else observed this behavior after upgrading to 26.9? Is there any additional configuration or update required for existing AI Skill objects to reflect AI Actions consumed?Thanks!

Hi Team, We are evaluating AI Skills/Generative AI usage in Appian and would like to monitor token consumption for optimization and cost governance purposes. I have a few questions: Is there a way to view the input tokens, output tokens, and total tokens consumed by an AI Skill execution? Does Appian expose token usage metrics at the AI Skill or process level? What is the recommended approach to track and analyze token usage for prompt optimization? Has anyone built a custom monitoring/dashboard solution to estimate token consumption for AI Skills and Agents? thanks in advance.

Hello community.I am currently working on the "Develop with AI: Get Started with Appian Composer" sub-course within the "Appian AI" course at the Appian Academy. As part of the hands-on exercise, I am instructed to create a request plan from the "Plan" tab and generate the application using Composer. Composer successfully creates all the objects, but when I access the "Build" section, I do not see the "Board" tab. I am using version 26.6 and am the application administrator. Does anyone know how to handle this? I need the "Board" tab to proceed with the exercise, as it opens the chat with Copilot AI, allowing me to interact with it.

any() 

Choosing the right topic can make the research process much easier. I usually look for a subject that has enough reliable sources, clear research questions, and room for discussion or analysis. Some students may consider whether to pay for research paper assistance, but I think understanding the research process and developing your own ideas is an important part of academic work.What factors do you consider when selecting a research topic?

The Appian environment language was switched from English to Spanish via the console. The problem is that label names have now disappeared from the processes, and the interfaces included in "User Input Tasks" have vanished. Why is this happening, and how can it be fixed—given that the language must remain Spanish (we cannot revert to English)? I have 90 processes; is there an automated way to handle this?

Hi Folks, I am using Appian AI Skill to extract the data from a pdf document. I uploaded multiple documents in the same batch and some of the documents are failing and giving the below-mentioned error.  Did anyone face the same issue or any insights into it? Your help is much appreciated.

Hi Everyone, I am facing an issue while creating an Extraction Model using AI Document Center. During the Extraction Model creation wizard (Steps 1–6), the final step displays the list of objects that will be created ( 4 AI Skills and 2 Process Models). After clicking “Confirm”, I receive a success message indicating that the Extraction Model has been created successfully. However, the underlying objects that are supposed to be generated are not actually created. I have verified this across the entire environment and cannot find the AI Skills or Process Models that were listed in the final step. Because these objects are missing, when I try to run the extraction from the UI, the process fails with an error indicating that the required Process Model cannot be found. A few additional observations from my debugging: This issue started occurring only recently. The same setup was working fine earlier. No application objects or configurations have been modified. The E

Yes, reaction time can be trained daily with consistent practice. Short sessions of 5 to 10 minutes using tools like the Human Benchmark test are effective. Daily repetition helps the brain adapt and respond faster to stimuli. However, overtraining can lead to fatigue, which may temporarily reduce performance. It is important to balance practice with rest and proper sleep. Incorporating different activities such as gaming, sports, or coordination drills can also enhance results. Over time, daily training can significantly improve both speed and consistency in reaction performance.

Hi Team, I am facing an issue while using AI Chat Agent. I have created a chat agent and plugged in with an interface rule. When I start with conversation, it has started and entered into “Thinking” status and not producing the response even though the instance is completed successfully (I checked it via Monitor tab). How to fix this issue? Is this issue specific to my appian enviornment (Current version is 26.6)?For one of the instances, I did not get the response even after waiting for more than 5 hours. Please refer to the attached screenshots for your reference. Note, Is there any possibility to test a chat agent before integrating with an interface rule? Currently, we able to test Process Agents before we integrate them with a process model.

Is AI Composer designed to generate a common application template regardless of the provided requirements?I've been experimenting with it using a relatively simple single-page application use case. However, instead of generating an application tailored to the specific requirements, it consistently produces a generic application template with standard pages and little to no relevant actions or functionality.Has anyone else experienced this behavior? Are there any best practices for prompting AI Composer to generate more use case-specific applications, or is the current expectation that it primarily creates a baseline template which needs to be refined manually afterward?

How to extract email attachments and email data by using AI

ALTFTool is a platform that helps users discover, explore, and compare alternatives to software, AI tools, and online utilities. Find the right tools based on your needs, features, and workflow