Copilot no longer supports image uploads
Copilot is no longer allowing image uploading/uploading. Neither in web mode nor in work mode. Does anyone know if they took away the resource for good or if it is some instability? The official documentation mentions that at least in web mode it should be supported. Document Reference: File formats supported by Copilot - Microsoft Support
Allowing Non-Technical Users to Update Topics/Responses
We have a Copilot Studio chatbot that is functioning well for the most part. However, the business has requested the ability to add and manage topics and responses to better address evolving needs. IT, however, is hesitant to grant non-technical users direct access to Copilot Studio, as it could risk breaking the chatbot. We've attempted to use a SharePoint list to manage topics and responses. This approach uses a Power Automate flow to access the SharePoint data and perform a simple lookup against user questions. Unfortunately, this solution lacks AI capabilities and feels more like a basic keyword-matching system than an intelligent chatbot. We’re exploring alternative solutions to allow non-technical users to manage topics and responses while preserving AI functionality and without giving direct access to Copilot Studio. Specific Questions: Are there ways to extend or customize Copilot Studio to allow controlled, user-friendly topic and response management? What other approaches or tools can we use to implement this requirement while maintaining the chatbot’s AI capabilities and ensuring IT governance? Any guidance, suggestions, or examples would be greatly appreciated!Chat Management in CoPilot desktop (browser) and Mobile
I love CoPilot, it's been my go-to assistant for months now. However I think a few features would contribute to it's long-term usability. Pin conversations. Keep certain recurring conversations at the forefront in the list of chats. This would be useful for daily chats such as a food diary, workout tracker, etc. Delete or Archive conversations. The ability to either delete individual conversations or archive/hide them to keep the chat history list clean, concise and relevant. Rename and Label conversations. As a daily CoPilot user I want the ability to name my chats and/or add labels to them so that I can easily find particular chats and add a bit of structure to my chat history list. Bonus Request - The ability to customize the start screen or bypass it altogether. I'd rather jump straight into my chats and use CoPilot for its core features. The suggested content in the beginning screen is simply an extra click and (for me) serves no practical purpose.
Change the default response from American English to British English in org level
The current response Enlish language is in American Language by default. I know we can ask Copilot to change the default response language from American Language to British Enlish. But wants to know if there is any way to achive the goal in Org level, instead of the end user proceed one by one.Planner Connector for Copilot
I added the planner connector for Copilot to be able to query my plans from BizChat. The connector shows it is installed in my Copilot alongside the Web Content connector. However, any prompts to Copilot regarding Planner plans and tasks are not answered. Can anyone provide guidance on how to use this connector or help me remove it from my Copilot.
Request to increase context length
Hello, I think users would greatly appreciate the ability to include more context in prompts. In my experience, results can be dramatically improved if you can include more examples and background information directly in the prompt. While you can attach longer files, this is in my experience not nearly as effective as directly including information in the prompt, as attaching files I assume uses something like RAG to pull snippets of your context instead of the entire thing. This is especially important when including examples as in few-shot prompting. If I want to use Copilot to help me write a longer document, I want to provide multiple already completed examples of this longer document. However, with the context limit of Copilot for M365 I might not even be able to fit one example in the prompt. This is not an issue with ChatGPT which has a much longer context length (128k tokens vs. 16k characters in Copilot). If users are able to write prompts with more context I think they will see much better results and therefore see the value of Copilot for M365 in a much clearer way. Thank you for your consideration.
Declarative Agent - Adaptive Card showing only "sometimes"
I am creating a declarative Agent using VS Code. I followed this tutorial: https://learn.microsoft.com/en-us/training/modules/copilot-declarative-agent-action-api-plugin-adaptive-cards-vsc/3-exercise-return-rich-responses-adaptive-cards Problem Moving on from the tutorial to my actual application, I see that the Adaptive Card is rendering only in 50% of cases (and it should be in 100% of cases). expected behavior I use the chatbot to make a POST request to my API. The response should be rendered in an adaptive card. relevant code elements of my app openapi.json : OpenAPI Specs (JSON) - please not I inserted a lot of "Something" for privacy reasons. { "openapi": "3.0.0", "info": { "title": "something", "version": "2", "description": "some description", "contact": { "name": "some name", "url": "some url" } }, "servers": [{ "url": "some server", "description": "some description" }], "paths": { "/chat/generate": { "post": { "summary": "something", "description": "Something", "operationId": "generateSomething", "parameters": [ { "name": "Something-1", "in": "header", "required": true, "description": "Something", "schema": { "type": "string" } }, { "name": "Something-2", "in": "header", "required": true, "description": "Something", "schema": { "type": "string" } } ], "requestBody": { "required": true, "content": { "application/json": { "schema": { "type": "object", "properties": { "v": { "type": "string", "description": "Specifies the API version being used.", "default": "2" }, "slidecode": { "type": "object", "properties": { "title": { "type": "string", "description": "Something" }, "subtitle": { "type": "string", "description": "Something" }, "slidenum": { "type": "integer", "description": "Something" }, "image_id": { "type": "string", "description": "Something" }, "body": { "type": "array", "items": { "type": "object", "properties": { "point": { "type": "string", "description": "Something" }, "description": { "type": "string", "description": "Something" }, "icon": { "type": "string", "description": "Something" } }, "required": ["point", "description", "icon"] } }, "talktrack": { "type": "string", "description": "Something" }, "sources": { "type": "array", "items": { "type": "object", "properties": { "title": { "type": "string", "description": "Something" }, "link": { "type": "string", "description": "Something" } }, "required": ["title", "link"] } }, "force_edit": { "type": "boolean", "description": "Something" } }, "required": [ "title", "subtitle", "slidenum", "image_id", "body", "talktrack", "sources", "force_edit" ] } }, "required": ["v", "slidecode"] } } } }, "responses": { "200": { "description": "Something", "content": { "application/json": { "schema": { "type": "object", "properties": { "data": { "type": "object", "properties": { "image_url": { "type": "string", "description": "Something" }, "presentation_view_url": { "type": "string", "description": "Something" } } } } } } } } } } } } } ai-plugin.json { "$schema": "https://developer.microsoft.com/json-schemas/copilot/plugin/v2.2/schema.json", "schema_version": "v2.2", "name_for_human": "Something", "description_for_human": "Something", "namespace": "Some-Namespace", "functions": [ { "name": "generateSomething", "description": "Something", "capabilities": { "response_semantics": { "data_path": "$.data", "properties": { "title": "$.presentation_view_url", "url": "$.image_url" }, "static_template": { "type": "AdaptiveCard", "$schema": "http://adaptivecards.io/schemas/adaptive-card.json", "version": "1.5", "body": [ { "type": "Container", "items": [ { "type": "Image", "url": "${image_url}", "size": "large" }, { "type": "TextBlock", "text": "${presentation_view_url}", "weight": "Bolder" } ] } ], "actions": [ { "type": "Action.OpenUrl", "title": "Get Something", "url": "${presentation_view_url}" } ] } } } } ], "runtimes": [ { "type": "OpenApi", "auth": { "type": "None" }, "spec": { "url": "apiSpecificationFile/openapi.json" }, "run_for_functions": ["generateSomething"] } ], "capabilities": { "conversation_starters": [ { "text": "Something" }, { "text": "Something 2" } ] } } instruction.txt # Copilot Agent Instructions ## Agent Identity You are doing SOMETHING ## Core Process ### Creation something - execute `POST /chat/generate` API call - Include the following required headers in the API call: 1. `something1`: something 2. `something2`: something - Ensure the API POST request includes the correct body (payload) with 1. the version `v` set to `2` 2. `slidecode` is an object with the following properties 2.1 `title` is something 2.2 `subtitle` is something 2.2 `slidenum` set to `1` 2.3 `image_id` set to `something` 2.4 `body` is a an array of 2 (two) objects, that have the required the properties `point`, `description` and `icon` - `icon` is a class name of Font Awesome Free 5.15 icon 2.5 `talktrack` is something 2.6 `sources` is a an array of 1 (one) object, that has the properties `title` and `link`, that both contain an empty string 2.7 `force_edit` is a boolen set to `true` - Example 1 for a `POST /chat/generate` body (payload) about something: ` json { "v": "2", "slidecode": { "title": "something", "subtitle": "something", "slidenum": "1", "image_id": "something", "body": [ { "point": "point", "description": "blub", "icon": "icon" }, { "point": "blub2", "description": "something", "icon": "icon" } ], "talktrack": "something", "sources": [{ "title": "", "link": "" }], "force_edit": "true" } } `
