Skip to main content

#Agentforce321 人がディスカッション中

Agentforce is an umbrella term of the set of tools to create and customize agents as well as the collection of agents Salesforce provides across the Customer 360. Agents can either be autonomous or assistive.

Hi everyone,  

I gave my first Salesforce Certified Administrator on 9th march 2026,  but i was unable to clear it at that time now i want to retake the exam but now I do not have 100 $ to spend on the retake how can i get full waiver or how can I get 100 $ voucher/ 200 $ voucher (If i can give a new attempt) 

 

#Free Vouchers  #Trailhead Challenges  #Trailhead  #Salesforce Admin  #Salesforce Developer  #Agentforce  #TrailblazerCommunity

3 件の回答
  1. 今日、17:02

    Hi @Mudit Chandwani,  

    there are several possible ways for you to try to get vouchers :  

     

    Take monthly trailhead quests to try to win gifts/swags and even vouchers : 

    https://trailhead.salesforce.com/quests

     

    If your company is Salesforce partner, you may have free vouchers (or discounted ones) according to your partnership level. See with your Salesforce's Partner Account Manager, or the person in charge in your company. 

    Good luck for your certification exam preparation ! 

0/9000

Hello Trailhead Support, 

 

I am working on the following Trailhead module: 

 

Module: Agent Customization with Flows 

Unit: Add a Flow as an Agent Action 

Trail: Become an Agentblazer Legend 2026 

 

I created a new Custom Developer Edition org using the specific signup link provided in the previous unit, Create an Agent-Ready Flow. 

 

I successfully enabled Einstein and Agentforce and completed the required flow challenges in this same org. However, when I open: 

 

App Launcher → Agentforce Studio → Agents 

 

the expected agent named: 

 

Coral Cloud Booking Agent 

 

is not available. 

 

The unit does not provide instructions to manually create this agent and appears to assume that it is automatically provisioned in the special Developer Edition org. 

 

I have already tried the following troubleshooting steps: 

 

Confirmed that Einstein is enabled. 

Confirmed that Agentforce is enabled. 

Turned Einstein off and on again. 

Refreshed the browser and logged into the org again. 

Checked both Agentforce Studio and Agentforce Agents in Setup. 

Confirmed that I created the org using the signup link provided specifically for this module. 

 

The Coral Cloud Booking Agent is still missing, which prevents me from completing the next hands-on challenge and adding my flows as agent actions. 

 

Could you please verify whether my Developer Edition org was provisioned correctly or provide instructions to restore or create the expected Coral Cloud Booking Agent? 

 

Organization username: [

feliperiquettoteste@gmail.com

] 

Organization creation date: July 24, 2026 

 

Thank you. 

 

#Trailhead Challenges

 

 

#Agentforce

3 件の回答
0/9000
1 件のコメント
0/9000

Hi everyone 👋

I’m testing Agentforce Voice in Enhanced Chat v2 and facing an issue with actions that use a Custom Lightning Type (CLT) output.

Here’s what I observed:

  • 🎙️ Voice conversation and basic responses work correctly.
  • ✅ An Apex action with String input + String output works correctly through voice.
  • ❌ When I use String input + CLT output, the voice interaction fails with messages like “I’m having trouble hearing you” or “I had some audio issues.”
  • 💬 The same action works correctly when tested through text.
  • ❌ I also tested CLT input + CLT output, and the same voice issue occurs.
  • 🌐 The issue happens in both Test Enhanced Web Chat and the deployed Experience Cloud site.
  • 🛠️ I simplified the sub-agent instructions and confirmed that the action works when CLTs are removed.

Has anyone experienced Agentforce Voice failing when an Apex action returns a Custom Lightning Type output in Enhanced Chat v2? 🤔

Is there any specific configuration or limitation for using Output CLTs with Agentforce Voice, or anything specific I should check in the CLT schema, renderer, or LWC?

Any suggestions or guidance would be really helpful. 🙏

Thank you! 😊 

#Agentforce

 

#Salesforce Developer 

2 件の回答
  1. 昨日、8:01

    The behavior points more toward a voice-channel serialization or rendering issue with the Custom Lightning Type rather than the Apex action itself. Since the same action works through text and succeeds when the CLT is removed, I’d isolate the CLT schema and test a simple structured output first, then add fields back one at a time. I’d also check the voice action payload and supported output types in Enhanced Chat v2. 

    For related troubleshooting discussions, community resources and practical solutions, these types of developer communities can also be useful references.

0/9000

I'm working on challenge 1 of the apex for agentforce superbadge. Its requesting to add the following actions:

  1. Lookup Contact by Email: This action queries the contact record based on an email address.
  2. Get Most Recent Booking by Contact: This action queries the most recent booking record before today.
  3. Lookup Experience: This action queries the experience record based on an experience name.

But these actions are not available in the org; Am I missing something? 

 

 

#Trailhead Superbadges  #Trailhead  #Agentforce

2 件の回答
  1. 昨日、2:43

    Hi Ebony, 

     

    Good news: these are pre-built actions that should already exist in the special superbadge org — you're not meant to build them in Challenge 1 (that's Exercise 2, where you create an Invocable Apex class). If they're not showing up, this points to one specific likely cause: you may not be using the special Developer Edition org that comes with this superbadge's pre-loaded Coral Cloud Resorts configuration and sample data. 

     

    To check/fix: 

    1. Confirm you signed up for the org through the "Sign up for a free Developer Edition org with special configuration" link on the superbadge unit page itself (Step I), not a generic playground or an existing DE org. This superbadge requires that specific special-config org — a regular org won't have these actions or the underlying Booking/Experience data at all. 

    2. If you already have a different org connected, disconnect it and sign up fresh via that link, then reconnect. 

     

    Once you're in the correct org, here's where to find them (per the official exercise walkthrough): in Agent Builder, when adding an action to your topic, use the Add Action search box: 

    - Search "Recent" → select Get Most Recent Bookings by Contact 

    - Search "Lookup" → select Lookup Experience 

    - (Lookup Contact by Email should appear the same way — search "Lookup" or "Contact") 

     

    These are found via search in that dialog, not browsed from a list — if you're clicking through a list instead of typing into the search field, that's worth double-checking too. 

     

    If you've confirmed you're in the correct special-config org and searching properly and they still don't appear, that would point to an org provisioning issue worth reporting via Trailhead Help with your org details. 

     

    Reference:

    https://trailhead.salesforce.com/content/learn/superbadges/superbadge-apex-for-agentforce

0/9000

Anyone else follow the instructions for getting started with Builder Central but don't have it showing up in setup? 

 

https://help.salesforce.com/s/articleView?language=en_US&id=platform.builder_central_set_up.htm&type=5

 

 

*On Enterprise with foundations and Einstein / Agentforce turned on. I have access to Customize Application and have tried both in Prod and a sandbox.  

 

 

 

#Salesforce  #Agentforce

1 件の回答
  1. 昨日、2:37

    Hi Josiah, 

     

    There are three prerequisites that are easy to miss even when Agentforce/Einstein look "on": 

     

    1. Agentforce entitlement specifically. Salesforce's own doc flags this separately from just having Agentforce features turned on — "Without the Agentforce entitlement, Builder Central won't load even when the beta setting is turned on." This is an org-level entitlement Salesforce provisions, not just a feature toggle you flip yourself. 

     

    2. Amazon Bedrock Models (Amazon and Anthropic) setting, in Einstein Setup. Builder Central's Build Agent specifically runs through Anthropic models via Amazon Bedrock — this is a separate checkbox from general Einstein enablement, and if your org restricts model providers, this is the one most likely to be missed. 

     

    3. Einstein Setup must be on — until it is, the doc says the Access Builder Central permission doesn't even appear as assignable, which would explain why nothing shows up for you. 

     

    The actual setup path is: Setup → Quick Find → "Builder Central" → Enable Builder Central (Beta), then assign the Access Builder Central permission via a permission set. 

     

    One gotcha specific to your sandbox test: if that sandbox was cloned from another, Einstein settings don't carry over in a clone — the doc explicitly calls this out, so if it's a cloned sandbox, re-enable Einstein Setup there and reassign the permission. 

     

    Given you've already got Customize Application and Agentforce/Einstein generally on in both Prod and a sandbox, but the Setup menu item itself isn't appearing at all (not just failing to load) — that points more toward #1: the Agentforce entitlement for Builder Central specifically may not be provisioned on your org yet, since it's still Beta and rollout isn't universal even for orgs with Agentforce otherwise enabled. Worth confirming with your Salesforce account team whether Builder Central Beta has been enabled for your org, since that's a Salesforce-side provisioning step, not something visible/toggleable from Setup alone if it hasn't been granted. 

     

    Reference:

    https://help.salesforce.com/s/articleView?language=en_US&id=platform.builder_central_set_up.htm&type=5

0/9000

Trailhead: Refine Your Prompt Template for Accurate Agent Replies. 

My agent is working in preview: 

 

Challenge not yet complete in Service Prompt Builder Whoops, looks like there was a problem. Please try again.

 

But still getting past the challenge. Any suggestion? 

 

#Trailhead Challenges  #Agentforce

3 件の回答
  1. 9月25日 6:34

    Hi Tuan, 

     

    I faced the same issue in the Trailhead challenge “Refine Your Prompt Template for Accurate Agent Replies.” 

     

    My Agentforce preview was working correctly, but Trailhead was still showing:

    “Challenge not yet complete. Whoops, looks like there was a problem. Please try again.” 

     

    I fixed it by checking the following steps: 

     

    1. Go to Setup → Flows.
    2. Search for Get Open Cases for Contact.
    3. Open the flow and make sure it is Active.
    4. Go to Setup → Prompt Builder.
    5. Open the prompt template Guest Experience Compensation.
    6. Make sure you are working on the latest version of the prompt template.
    7. In the prompt, after Case Details:, insert: Insert Resource → Flow → Get Open Cases for Contact
    8. After Complaint:, insert: Insert Resource → Free Text → Complaint
    9. Do not manually type the resource names. Insert them using the Insert Resource option.
    10. Save the prompt template.
    11. Test it in Preview with Sofia Rodriguez.
    12. Confirm that the resolved prompt contains:
      • Contact details
      • Open case details
      • Complaint
      • Compensation instructions
    13. Click Activate for the latest version of the Guest Experience Compensation prompt template.
    14. Go back to the Coral Cloud Experience Agent.
    15. Refresh the agent preview and test again.
    16. Use: The White Water Rafting experience my family and I booked was unexpectedly cancelled.
    17. When the agent asks for customer details, enter:  I'm sofiarodriguez@example.com and my membership number is 10008155.
    18. If the agent response is working correctly, go back to Trailhead and click Check Challenge again.

     

    The main things to verify are: 

     

    Get Open Cases for Contact Flow → Active

    Guest Experience Compensation → Latest Version Active

    Case Details → Flow resource inserted correctly

    Complaint → Free Text resource inserted correctly

     

    Thank You, 

    Venkat Yadav.

0/9000

 

Need an Solution please help !

 

 

I am currently working on the Agenblazer Innovator status 2026

 

 

 

While working on the

Agentforce Grid -> Explore Data with Agentforce Grid

 

module i am facing this error eventhough the every configurations are right. Is any agentforce expertice people it would be great to help me!  

  

  

[LXM001] A temporary system error occurred. Please provide this request ID to customer support. RunId: 01a0d49a-4f08-77ec-ad16-74a1a4242732  

  

Thanks in Advance!  

 

 

#Agentforce  #Trailhead  #Trailhead Challenges  #Salesforce Developer

1 件の回答
  1. 9月25日 13:36

    I've just completed this module and had a similar thing happen first time I tried. If right-click on the column and click run hopefully it will resolve the issue it did for me.

0/9000

Hi everyone, 

 

I'm setting up Financial Advisor Assistance in Agentforce for Financial Services and I'm running into an error when trying to activate the agent.

Here’s what I've done so far:

  1. In Setup → Agentforce for Financial Services, I enabled Financial Advisor Assistance.
  2. I created an agent using the Financial Advisor Assistance agent template.
  3. When I try to activate the agent, I get the following error:   

       

    "An error occurred when processing the <Entity> entity with name Financial_Advisor_Assistance. Looks like you don’t have access to agent templates associated with the AgentforceEmployeeAgent agent type. Make sure that you have the required permissions for the type and try again."  
  4. Financial Advisor Assistance agent activation error - AgentforceEmployeeAgent permissions     

         

    The confusing part is that I cannot find any permission set or permission like

    AgentforceEmployeeAgent in my org.     

         

    According to the Salesforce documentation, Financial Advisor Assistance requires permissions such as

    Einstein for Financial Services, Financial Advisor Assistance, Agent Platform Builder, and Execute Prompt Templates, but I'm not sure if there is an additional permission/license required.     

         

    Has anyone encountered this specific error when activating the

    Financial Advisor Assistance agent?     

        

     

    • Is there any other setup required before activating the Financial Advisor Assistance agent?
    • Is this possibly an issue with the availability/access of the Financial Advisor Assistance employee-agent template in a pre-release FSC trial org?

 

Any guidance on how to resolve this would be greatly appreciated. 

Thanks!

#Financial Services Cloud  #Financial Advisory  #Agentforce @* Financial Services Cloud *

1 件の回答
  1. 9月25日 10:18

    Hi Jayashree, 

     

    Creating/activating an Agentforce Employee Agent needs two specific user permissions you haven't listed — 

     

    To create an Agentforce Employee agent: 

    - Manage AI Agents 

    - Manage Agentforce Employee Agents 

     

    These are separate from Agent Platform Builder and Execute Prompt Templates, which only cover general Agentforce/prompt usage — not specifically the AgentforceEmployeeAgent template type. Your error message text ("access to agent templates associated with the AgentforceEmployeeAgent agent type") maps directly to these two permissions being missing. 

     

    To fix: 

    1. Setup → Permission Sets → check if your assigned permission set(s) include Manage AI Agents and Manage Agentforce Employee Agents as user permissions (search under App Permissions or AI-related permissions in the permission set editor) 

    2. If missing, either add them to your existing permission set or create a new one with both checked, and assign it to your user 

    3. Retry the Commit/Activate action on the agent 

     

    This is likely why the error is specific to activation and not the earlier steps (enabling the feature, creating from template) — those work with Agent Platform Builder alone, but committing/activating an Employee Agent specifically checks for the two Manage permissions above. 

     

    Reference:

    https://help.salesforce.com/apex/HTViewHelpDoc?id=ind.insurance_agent_create_from_template.htm

0/9000

Hi Salesforce Community,

We are currently working on an Agentforce-based Knowledge Article generation use case and would appreciate recommendations from anyone who has implemented something similar.

Our objective is to automatically generate accurate, reusable, customer-facing Knowledge Article drafts from resolved Salesforce Cases, with the eventual goal of using those articles for customer self-service and case deflection.

Our current implementation uses Salesforce Prompt Builder and Agentforce capabilities. The Case information is supplied as grounding context, and the generated Knowledge Article is primarily focused on:

Title

 

Summary

 

Content / Resolution

The Resolution is especially important for us. We want the generated article to accurately represent the actual outcome of the Case and distinguish whether it resulted in a:

  •  Final Solution 
  •  Workaround 
  •  Improvement 
  •  Enhancement / Feature Request 

At the same time, the model should remain strictly grounded in the Case information and should not infer or introduce unsupported technical details.

One of the main challenges we are currently investigating is how to provide the model with the complete and relevant Case context.

Important information can be distributed across several Salesforce sources, including:

  •  Case Subject and Description 
  •  Case Comments 
  •  Email Messages 
  •  Work Notes 
  •  Resolution Notes 
  •  Related records or conversations 
  •  Other Case-related supporting information 

We have observed situations where useful information exists in the Case context but is not fully reflected in the generated Knowledge Article.

We also want to prevent the opposite behavior, where the generated article introduces resolution steps, troubleshooting, or technical information that is not actually supported by the Case.

We are therefore looking for guidance on the Salesforce-recommended or out-of-the-box approach for this type of implementation.

Specifically, we would appreciate input on the following:

  1.  Is Einstein Knowledge Creation / the standard “Create Knowledge Articles From Case” Prompt Builder template the recommended approach for automatically generating Knowledge Articles from resolved Cases? 
  2.  What is the recommended way to provide the complete Case context, including related information such as Case Comments and Email Messages, to a Knowledge Generation prompt without manually combining everything into one large input? 
  3.  For a Knowledge Generation Prompt Template using Case as the grounding record, what additional Salesforce-supported data sources or related resources are typically recommended? 
  4.  Is there an out-of-the-box Salesforce approach for reliably identifying and generating the actual resolution from a resolved Case while keeping the output strictly grounded? 
  5. Has anyone implemented prompt logic that successfully distinguishes between:
    •  Final Solution 
    •  Workaround 
    •  Improvement 
    •  Enhancement / Feature Request 
  6. without causing the LLM to over-infer or classify an unsupported outcome?
  7.  What are the recommended Prompt Builder techniques for ensuring that generated Knowledge Articles remain: 
    •  Accurate 
    •  Grounded 
    •  Customer-friendly 
    •  Reusable 
    •  Resolution-focused 
    •  Suitable for customer self-service and case deflection 
  8.  If the standard Knowledge Creation functionality is not sufficient for this use case, are there recommended approaches using Flow, Prompt Builder, Agentforce, Data 360, related record grounding, or Retrievers that work well for generating complete Knowledge Articles from Salesforce Cases? 
  9.  Are there any existing Salesforce templates, accelerators, AppExchange solutions, or recommended implementation patterns that could help avoid building a heavily customized solution from scratch? 

We are particularly interested in hearing from anyone who has implemented a production workflow similar to:

Resolved Salesforce Case → Grounded Case Context → AI-Generated Knowledge Article → Agentforce / Customer Self-Service

Any implementation examples, best practices, known limitations, configuration recommendations, or workarounds would be greatly appreciated. 

We would also appreciate any

out-of-the-box approaches, alternative implementation patterns  beyond the standard Salesforce capabilities currently available.

If anyone has successfully implemented a different approach for generating high-quality, grounded Knowledge Articles from Cases, we would be very interested in learning about the architecture, tools, or configuration used

Thank you in advance for your guidance. 

 

#Agentforce  #Salesforce  #TrailblazerCommunity  #Trailhead Challenges

1 件の回答
  1. 9月25日 10:17

    Hi Bijay, 

     

    On Q1: this has actually changed. Einstein Knowledge Creation now supports custom Prompt Templates via the "Knowledge Generation" type in Prompt Builder — not limited to the single default case-only template anymore. Build a custom template tailored to your use case; that's the current recommended path. 

     

    For pulling in Comments/Emails/Work Notes without manual concatenation: 

    - Flow-based grounding: attach a Flow as a grounding resource in Prompt Builder — it queries related records and feeds them in structured, instead of one giant merge field. 

    - Retrievers (Data Cloud): for larger/unstructured volumes, index content into a Data Library and use a Retriever (via Einstein Studio) for semantic retrieval instead of dumping everything into one flow output. 

     

    For grounding accuracy / avoiding invented resolution steps: 

    - Add an explicit negative instruction in the prompt: "Only use information present in the provided context. Do not infer steps not explicitly stated." This measurably reduces over-inference. 

    - Split classification from generation — use a separate Field Generation prompt just to classify Final Solution/Workaround/Improvement/Enhancement (same context), then pass that result into the Knowledge Generation prompt. Two narrow prompts outperform one prompt doing both. 

    - Add 2–3 labeled few-shot examples of each resolution category in the prompt instructions — this taxonomy is business-specific, so there's no built-in classifier; examples are what actually stabilize classification. 

     

    No native accelerator exists for the resolution-type classification specifically — that part is custom prompt engineering, not a toggle. 

     

    Reference:

    https://help.salesforce.com/s/articleView?language=en_US&id=release-notes.rn_knowledge_creation_custom_prompts.htm&release=264&type=5

0/9000