Product feedback | The place for Zendesk users to come together and share
Skip to main content

Filter by feedback status

Filter by product

273 Requests

Eric134Newcomer

PRODUCT FEEDBACK: Transparent failure reporting and post-failure state for AI-built action flowsFeedback submitted

When the AI action flow builder fails to create a flow — which happens often on more complex, multi-step requests — it returns only a generic “Something went wrong.” There is no indication of what failed, why, which step or action caused it, or what category of failure occurred (the request could not be parsed, an action is unsupported, validation failed, the build timed out, a limit was reached). Equally, there is no indication of the resulting state: the user cannot tell whether nothing was created, or whether a partial, incomplete flow was left behind in the instance. Because the message is opaque and the post-failure state is invisible, it is not possible to understand the failure, correct the request, or safely retry it. On an admin surface that modifies live automation configuration, this is a governance concern as much as a usability one.Requested EnhancementSpecific, actionable error reporting: state what failed and why — the failing step or action, the specific constraint or limit involved, and the error category (parse failure, unsupported action, validation error, timeout, cap reached). If possible, include a reference ID and timestamp so the failure can be traced and reported to Support. The outcome must be unambiguous. Business ValueLets administrators diagnose and correct a failed request instead of guessing, turning a dead end into a fixable one.Prevents orphaned or partial automation objects from accumulating silently in a production instance.Provides a traceable reference for Support and internal escalation when a failure recurs.Protects configuration integrity in regulated, multi-brand instances, where an unnoticed partial flow could later fire, or be mistaken for a complete one.Builds the trust needed to adopt the feature beyond the simplest single-step flows.Current WorkaroundNone effective. The only options are to reduce the request to a single step until it succeeds, or abandon the AI builder and author the trigger/automation manually. Neither reveals what failed, and neither confirms whether a partial flow was left behind.Use CaseWe evaluate and operate in a regulated, multi-brand environment. Complex builds — multiple steps or multiple cadences in one request — currently frequently fail with this generic message, so we cannot tell whether the request was understood, which part exceeded the builder’s current capability, or whether a partial flow now sits in the instance. For a regulated operator that must account for every automation object and demonstrate configuration control, an opaque failure with an unknown resulting state is not adoptable. We need to see what went wrong, and be certain of the state the instance is left in afterwards.  

Feedback: Do not enable a feature thats going to break after 5 uses by design.Under review

Please give a quick overview of your product feature request or feedback and note who in your org is affected by this issue [ex. agents, admins, customers, etc.]. (2-3 sentences) This feedback concerns a recent announcement about the Copilot AI writing tools.Introducing Copilot AI writing tools for Professional plans and above I don't want this enabled by default. I don't want to enable a feature in my environment that's going to stop working after a limited number of uses. This should be DISABLED by default. What problem do you see this solving? (1-2 sentences) I currently do not have the Copilot addon, so I do not have a way to pre-emptively disable or limit this feature to specific groups ahead of time. The announcement says to go to this article to find instructions, but the instructions can't be followed until the feature is already rolled out to my users. I'm guessing I'll have the option to disable this after the feature is enabled in my instance, but that's not really clear. What would be your ideal solution to this problem? How would it work or function? (1-2 sentences)Don't enable features by default that will break after a limited number of uses.  From what I can tell, this is just Grammarly, but only for Zendesk. We already have Grammarly, which works across all applications. Why would I want this?  I don't want something like this enabled by default, then break after x number of comments, and then have 240 support agents Slacking me about this error.This message efficiently says “Slack Chris when you see this error”. 

Voice AI agents: let "speak to a person" run a procedure (callback intake) when there is no live phone teamFeedback submitted

What is your request? Who does it affect in your organization? (2–3 sentences)We'd like voice AI agents to let admins choose what happens when a caller asks for a person: transfer to the phone queue, or run a procedure (for example, collect callback details and create a routed ticket). We'd also like a way to turn off or edit the generated hand-off lines. This affects our callers, our area operations and safety teams who handle callbacks, and our admins.What problem would this solve? (1–2 sentences)We have no live phone team, so every request for a person ends in voicemail after the AI agent tells the caller "I'll get someone on the line for you." Callers hear a promise we can't keep, and their request arrives without the structured name, address and contact details the AI agent collects on every other path.When did this issue last affect you? How often, and what impact? (3–4 sentences)It happens daily: in our first week live (November 4 to 10, 2026), 6 of 16 calls opened by asking for an agent, manager or supervisor, and 8 of 16 ended in voicemail. Each voicemail needs manual listening and triage before it reaches the right area team, because no name, address or request type is captured. On launch day, five real callers' tickets reached no team at all until we added routing rules. Callers also hear a contradictory "I'll connect you" followed by "Sorry, I can't connect you," which hurts trust in the assistant.Are you using any workarounds? (1–2 sentences)We use an Escalation block that sends callers to voicemail with honest wording, a greeting that offers to take a message, and triggers and automations that assign every voicemail ticket to a team. We tried a "Link to" use case in the Escalation reply, but the AI agent stayed in hand-off mode, invented transfer promises and never ran the linked use case.What would your ideal solution look like? (1–2 sentences)A per-agent setting such as "When a caller asks for a person: transfer to phone queue | run a use case or procedure," plus a toggle to disable or customize the generated hand-off lines on voice. Ideally, escalations that end in voicemail would also be labeled as assisted escalations in AI agent analytics.  

Add the current ticket ID to the Get ticket info actionFeedback submitted

In AI agents, the Zendesk Support action Get ticket info can return the assignee ID, the group ID, and the problem ticket ID, but not the ID of the ticket the conversation is already running on. This affects admins who build procedures and any external integration that must update that exact ticket.What problem do you see this solving?An integration could receive the current ticket ID as a normal action parameter, in the same way it already receives other ticket fields. We would not need a custom field, a trigger, and a webhook just to learn which ticket the AI agent is handling.When was the last time you were affected by this? What happened? How often does this occur, and how does it impact your business?This happened today, 9 November 2026, on a live invoice request.Our AI agent handles "send me the invoice for my order" by calling an n8n workflow. The workflow downloads the invoice PDF and attaches it to the Zendesk ticket. The order number is not enough to find the ticket: several tickets can share the same marketplace order ID. The workflow has to update the ticket the customer is writing on, so it needs that ticket's ID.Get ticket info does not offer the current ticket ID. The closest field is Problem ticket ID, which is a different ticket and is empty on a normal request. We passed that value into the API and the integration received ticket_id: 0. The invoice was found, but it could not be attached to the right ticket.Are you currently using a workaround?Yes. We created a custom text field that agents are not meant to edit. A trigger runs when the field is empty and calls a webhook. The webhook sends PUT /api/v2/tickets/{{ticket.id}}.json and writes {{ticket.id}} into that custom field. The AI agent then uses Get ticket info on the custom field and passes the value to n8n.A trigger action cannot write {{ticket.id}} into a text field directly. Zendesk stores the placeholder as literal text, so the webhook is required. The webhook runs a few seconds after the ticket update, so the AI agent can still call n8n before the field is filled.What would be your ideal solution? How would it work?Add Ticket ID to Get ticket info, next to Assignee ID and Group ID. The action would save the ID of the current ticket as a parameter, and that parameter could be sent to an API integration. No custom field, trigger, or webhook would be required, and the value would be available in the same step as the other ticket fields.

Allow agents to view auto assist procedure sources without Knowledge admin accessFeedback submitted

Feature requestToday, the source of an auto assist suggestion is only shown to agents who have permission to view that specific procedure. In practice, this means only Knowledge admins can see which procedure was used. Agents without Knowledge admin access only see the rationale and a generic “Learn more about other sources” link.Our use caseWe have several published auto assist procedures for our Support team. Our team leaders review auto assist suggestions daily and help us identify inconsistencies so our Knowledge team can improve the procedures.When a leader sees a suggestion, they can read the rationale (for example, “The procedure states that...”), but they can’t see which procedure was used. To find out, they have to escalate to an admin, who reproduces the scenario. We confirmed with Zendesk Support that this is expected behavior and that there’s currently no separate permission for this.The only way to give them visibility today is to enable “Manage knowledge” on their role. That grants full Knowledge admin access, including editing and publishing articles and procedures, which is much more than they need.What we’d likeA view-only permission (for example, “View procedure sources in auto assist”) that lets agents in custom roles:see the title of the procedure used in a suggestion; open the procedure in read-only mode;without being able to create, edit, publish or delete knowledge content.Why it mattersLeaders and senior agents can give precise feedback on procedures, which improves auto assist quality over time. Admins don’t need to reproduce tickets just to identify the source. It keeps the principle of least privilege, since we wouldn’t have to grant admin access only for visibility.

Unified Status Permissions for Non Admins, such as Managers or Leads For App BuilderFeedback submitted

Please give a quick overview of your product feature request or feedback and note who in your organization is affected by this issue [ex. agents, admins, customers, etc.] (2-3 sentences)Hello, I’ve created the perfect app that replaces the Precanned Explore Dashboard “Live Data” within ZenDesk that allows us to change our team member’s unified status within ZenDesk, however only perfect to admins, as only admins is allowed to change Unified Statuses from an API standpoint. What problem do you see this solving? (1-2 sentences)The Precanned Explore Dashboard “Live Data” Specific roles are allowed to change Unified Statuses, there’s a setting that allows non admins to do so (including non admin roles) When was the last time you were affected by this lack of functionality, or specific tool? What happened? How often does this problem occur and how does this impact your business? (3-4 sentences)When our team members have outages or loses connection and need their managers to switch their unified statuses; this explore dashboard can be clunky as it’s not apart of ZenDesk Support and has to be found in a separate page and on top of that it’s buried in a analytical dashboard rather than front facing for our support managers. Are you currently using a workaround to solve this problem? (If yes, please explain) (1-2 sentences)No workaround just using the dashboard What would be your ideal solution to this problem? How would it work or function? (1-2 sentences)ZenDesk allowing custom role perms to allow those in the role to change their team member’s Unified Status, I don’t completely understand why they can’t have the access to do so in a custom built ZenDesk app if they can change statuses from the Explore Dashboard. 

Shaun11Contributor

Real-time AI Suggestions for Voice Calls Should Use Auto-Assist ProceduresFeedback submitted

Please give a quick overview of your product feature request or feedback and note who in your organization is affected by this issue [ex. agents, admins, customers, etc.] (2-3 sentences)We would like Zendesk’s real-time AI suggestions for voice calls to support Auto assist procedures as a source, not only help center content. This affects agents who rely on live suggestions during customer calls, admins who maintain approved procedures, and customers who need accurate, complete guidance during live support interactions. Zendesk’s current documentation shows that real-time voice suggestions search help center content, while Auto assist procedures are a separate feature and cannot currently be configured on voice channels. [support.zendesk.com], [support.zendesk.com], [support.zendesk.com]What problem do you see this solving? (1-2 sentences)Real-time voice suggestions can paraphrase or simplify information from help center content instead of following our exact operational procedures. Allowing voice suggestions to use Auto assist procedures would help agents receive step-by-step, admin-approved guidance during calls, improving consistency and reducing the chance that important procedural details are missed. [support.zendesk.com], [support.zendesk.com]When was the last time you were affected by this lack of functionality, or specific tool? What happened? How often does this problem occur and how does this impact your business? (3-4 sentences)We are affected by this problem every time an Agent uses Real-time AI Suggetions in a call. This problem is especially important for voice calls because agents need concise, accurate guidance in real time and do not have the same opportunity to review long articles while speaking with a customer as they do during Messaging or Email tickets. Zendesk’s voice suggestion documentation currently describes help-center-based suggestions, while the Auto assist documentation says procedures are the primary knowledge source for Auto assist and that Auto assist cannot be configured on voice. [support.zendesk.com], [support.zendesk.com], [support.zendesk.com]Are you currently using a workaround to solve this problem? (If yes, please explain) (1-2 sentences)Yes. We are trying to maintain help center content and procedures separately, but that creates a gap: real-time voice suggestions rely on help center content, while our more precise step-by-step guidance lives in Auto assist procedures. Zendesk’s documentation says procedures can include macros, actions, agent instructions, help center content, other procedures, and ticket fields, but the real-time voice suggestion documentation does not indicate that those procedures can be used as a source for live voice suggestions. [support.zendesk.com], [support.zendesk.com]What would be your ideal solution to this problem? How would it work or function? (1-2 sentences)Ideally, Zendesk would allow admins to enable Auto assist procedures as a source for real-time AI suggestions during voice calls, either globally or by group, line, brand, intent, or call reason. During a live call, Copilot should use the real-time transcript to identify the relevant procedure, then surface the procedure’s approved steps, macros, actions, and agent instructions in the Knowledge or Copilot panel without requiring the agent to leave the call workflow. [support.zendesk.com], [support.zendesk.com]

Richard23Newcomer

API to create and maintain Zendesk AI agentsUnder review

Please give a quick overview of your product feature request or feedback and note who in your organization is affected by this issue [ex. agents, admins, customers, etc.] (2-3 sentences)We would like a supported API for creating and maintaining Zendesk AI agents, including AI agents Advanced. This affects our Zendesk administrators and Support team members responsible for building, reviewing and maintaining AI-agent workflows.What problem do you see this solving? (1-2 sentences)API access would allow us to automate agent configuration, review changes and maintain consistent behaviour across agents. It would reduce repetitive manual work and support a controlled, repeatable development process.When was the last time you were affected by this lack of functionality, or specific tool? What happened? How often does this problem occur and how does this impact your business? (3-4 sentences)We encountered this while developing a workflow to create and update test AI agents. We attempted to use browser automation to create agents and maintain their instructions and procedures. This limitation affects the process whenever we need to create an agent or update its configuration. It adds development and maintenance effort, slowing our ability to test and introduce new support workflows.Are you currently using a workaround to solve this problem? (If yes, please explain) (1-2 sentences)We have attempted browser automation as a workaround for creating and updating agents through the Zendesk interface. However, this creates its own maintenance overhead, as changes to the interface can break the automation and require further adjustments and testing.What would be your ideal solution to this problem? How would it work or function? (1-2 sentences)A documented, supported management API would let us create, clone, read and update agents, including their instructions, procedures, flows and configuration. It would also support exporting and importing definitions, testing changes and publishing approved versions, with version history and rollback.

Add Seconds in Wait block option in Custom ActionsFeedback submitted

Please give a quick overview of your product feature request or feedback and note who in your organization is affected by this issue [ex. agents, admins, customers, etc.] (2-3 sentences)We request adding "Seconds" as a time unit option to the Wait block in Zendesk Bot Builder alongside minutes, hours, and days. This directly impacts our bot admins who design automated conversation flows, as well as our end users who expect timely responses in chat. Currently, administrators cannot build brief pauses into custom action flows to accommodate short backend processing times.What problem do you see this solving? (1-2 sentences)It bridges the gap between instantaneous actions and long pauses by allowing the bot to wait briefly for an asynchronous worker to return data. This prevents premature fallback errors and eliminates unnatural 60-second delays in live chat conversations.When was the last time you were affected by this lack of functionality, or specific tool? What happened? How often does this problem occur and how does this impact your business? (3-4 sentences)We encountered this issue this week while deploying an automated custom action connected to an external worker that takes roughly 3 to 5 seconds to process user data. Because the minimum available wait duration is one minute, users either faced an unacceptably long silence or the bot moved to the next step before the data was ready, triggering failure messages. This impacts every customer who uses this automated flow across hundreds of conversations per day. The resulting drop in deflection rates frustrates customers and causes unnecessary ticket escalations to human agents.Are you currently using a workaround to solve this problem? (If yes, please explain) (1-2 sentences)Yes, we currently force a minimum 1-minute wait or rely on synchronous API calls that frequently risk timeouts. Both options result in high user drop-off and a degraded messaging experience.What would be your ideal solution to this problem? How would it work or function? (1-2 sentences)The ideal solution is adding "Seconds" to the duration unit dropdown in the Wait step, supporting values from 1 to 59. The bot would pause execution for the designated number of seconds before executing the subsequent action or polling step.  

Cloudflare restrictions prevent technical SEO audits and block AI crawlersFeedback submitted

We are experiencing limitations with our Zendesk-hosted Help Center that prevent us from properly managing its technical SEO health and ensuring our public content is accessible to AI-powered search engines.There are two major issues:1. Inability to perform technical SEO audits using third-party crawlersWe need to be able to crawl our public Help Center using third-party SEO tools such as Screaming Frog and Sitebulb. However, requests from these tools appear to be blocked by Zendesk's security infrastructure, preventing us from performing comprehensive technical SEO audits.This means we cannot reliably identify and address important technical SEO issues, including broken internal links, orphaned articles, missing H1 headings, incorrect canonical tags, indexability problems, and redirect issues.As a result, our SEO team cannot independently monitor the technical health of our Help Center or proactively identify issues that could negatively affect organic visibility and user experience.2. Uncertainty and restrictions around AI crawler accessibilityWe also need our public Help Center content to be accessible to AI-powered search engines and answer engines, including ChatGPT and Perplexity.With the increasing importance of AI-driven content discovery, it is essential that legitimate AI crawlers, such as OpenAI's OAI-SearchBot and PerplexityBot, can access, retrieve, and reference our public articles.However, we currently lack transparency regarding which AI crawlers Zendesk allows or blocks. If legitimate crawlers are blocked by Zendesk's security layer, our content may be excluded from these platforms, limiting its discoverability and visibility in AI-generated answers.What we need from ZendeskWe need Zendesk to provide a supported solution that allows customers to: Run third-party technical SEO crawlers against their public Help Centers without being blocked by Zendesk's security infrastructure. Allow legitimate AI crawlers to access public Help Center content. Configure or manage crawler permissions through supported security settings or allowlisting mechanisms. Understand which crawlers are permitted or blocked and what restrictions apply. Alternatively, Zendesk should provide built-in technical SEO auditing and monitoring capabilities that allow customers to identify and address technical issues independently.These limitations directly affect our ability to maintain our Help Center's technical SEO health and ensure our content remains discoverable across both traditional search engines and emerging AI-powered platforms.We believe customers should have the ability to independently audit their public Help Centers and control access for legitimate search and AI crawlers, without compromising Zendesk's security protections.  

Create a translate feature under the Enhanced writing optionsFeedback submitted

Please give a quick overview of your product feature request or feedback and note who in your organization is affected by this issue [ex. agents, admins, customers, etc.] (2-3 sentences)We’d like a built-in translation option added to the enhanced writing options, so agents can translate a drafted message into another language without leaving the editor. This especially is useful for Copilot generated answers. What problem do you see this solving? (1-2 sentences)Agents can’t reliably communicate with customers in languages they don’t speak, which slows responses and risks errors or inconsistent tone. Translation needs to happen in the same place where they already refine their writing. When was the last time you were affected by this lack of functionality, or specific tool? What happened? How often does this problem occur and how does this impact your business? (3-4 sentences)This week, an agent received a ticket in a language they don’t speak and had to leave the tool to translate both the customer’s message and their reply. It happens several times a week, and more often during peak periods. The extra steps increase handling time, delay first responses, and create a risk of mistranslations or inconsistent tone. This ultimately hurts customer satisfaction and agent efficiency. Are you currently using a workaround to solve this problem? (If yes, please explain) (1-2 sentences)Yes, agents copy text into external tools, then paste the result back. This is slow, breaks their workflow, and raises data privacy concerns when customer information is pasted into third-party tools. What would be your ideal solution to this problem? How would it work or function? (1-2 sentences)A “Translate” option under enhanced writing options that lets agents pick a target language and translate their draft in one click, preserving tone and formatting, with the option to review before sending. Ideally it would also auto-detect the language and suggest it as the default. 

Madalina11
Madalina11Newcomer

Automatic translations for embedded templates in Advanced AIFeedback submitted

Please give a quick overview of your product feature request or feedback and note who in your organization is affected by this issue [ex. agents, admins, customers, etc.] (2-3 sentences)The embedded templates supported by the Advanced AI do not currently support automatic translations, compared to the Essentials bot builder. I would like to suggest for this feature to be added for embedded templates. What problem do you see this solving? (1-2 sentences)Withe the Essentials bot builder we only had to create the flow in English and could rely on the automatic translations for all the languages we support in. We have two brands, one supporting in 16 languages and another one supporting in 39 languages. While I understand that some industries may wish to use manual translations due to sensitive terms being used, I find that removing completely the automatic translations from this feature is impacting a lot of customers who could actually benefit from this in the embedded templates.We could make a lot of use of these embedded templates, but are currently limiting our workflows to use as few as possible because manual translations are time consuming especially in our case. Not to mention that whenever we make a small change in the default English template this has to be manually changed for each language individually.  When was the last time you were affected by this lack of functionality, or specific tool? What happened? How often does this problem occur and how does this impact your business? (3-4 sentences)The embedded templates have been released without automatic translation and impacted us as soon as we started migrating to Advanced AI as our planning did not take into account time needed and costs for manual translations. Are you currently using a workaround to solve this problem? (If yes, please explain) (1-2 sentences)As a workaround, we are trying to avoid using templates and have only created them if absolutely necessary. What would be your ideal solution to this problem? How would it work or function? (1-2 sentences)Ideally, the templates should support automatic translations with the possibility to add manual ones for customers who might require precise translations for more sensitive areas as health or IT services.This would allow all customers to benefit from using templates in the best way suitable for their business. 

Elsa12
Elsa12Newcomer

Filter which users the AI agent responds to by tag or segmentFeedback submitted

Please give a quick overview of your product feature request or feedback and note who in your organization is affected by this issue [ex. agents, admins, customers, etc.] (2-3 sentences)We need a way to exclude specific users from the AI agent (Advanced) on a given channel, based on criteria such as user tags, user segments, organization or custom user fields. For example, on WhatsApp we want customers with a specific tag to skip the AI agent entirely and go straight to a human agent. This affects our admins, our agents, and the customers who should never be handled by the bot. What problem do you see this solving? (1-2 sentences)As an intermediary between customers and ticket and activity providers, some of the people contacting us need a human from the first message, and an automated conversation creates friction for them. Being able to filter who the AI agent responds to would let us keep the bot for the general public while protecting these specific relationships. When was the last time you were affected by this lack of functionality, or specific tool? What happened? How often does this problem occur and how does this impact your business? (3-4 sentences)This happens every time one of these tagged users contacts us through WhatsApp, because the AI agent activates for everyone on the channel with no way to filter. They have to go through the bot before reaching an agent, which delays resolution and gives a poor experience to contacts we want to treat with priority. Since we can’t segment the AI agent, we are forced to choose between deploying it on the whole channel or not using it there at all.  Are you currently using a workaround to solve this problem? (If yes, please explain) (1-2 sentences)No, we haven’t found a reliable workaround, since the AI agent’s activation can’t be conditioned on user data. What would be your ideal solution to this problem? How would it work or function? (1-2 sentences)An audience or eligibility setting per AI agent and per channel, where admins can include or exclude users based on tags, user segments, organizations or custom fields. Excluded users would bypass the AI agent and be routed directly to a human agent. 

Elsa12
Elsa12Newcomer

Allow removing or customizing the “automatically generated using AI” footer on AI agent replies, per channelFeedback submitted

Please give a quick overview of your product feature request or feedback and note who in your organization is affected by this issue [ex. agents, admins, customers, etc.] (2-3 sentences)We would like to user an AI agent (Advanced) to reply to customer reviews that arrive through the Trustpilot channel integration, and every reply it publishes ends with a fixed footer: “This email was automatically generated using AI.” There is no setting at account, agent or channel level to remove or customize it, and “Display sources for generative replies” is disabled. This affects our admins, who can’t control the content of public replies, and our customers and prospective customers, who read these replies on Trustpilot. What problem do you see this solving? (1-2 sentences)It would let us publish AI-generated review replies with wording that is accurate for the channel and consistent with our brand. Today the footer refers to an “email” on a public review platform and stays in English regardless of the reply’s language. When was the last time you were affected by this lack of functionality, or specific tool? What happened? How often does this problem occur and how does this impact your business? (3-4 sentences)It happens on every single reply the AI agent publishes on Trustpilot, so we are affected daily. Each reply appears publicly with a footer that calls it an email, which looks like an error to anyone reading our reviews. Since review replies are one of the most visible parts of our brand to potential customers, this undermines the quality of our public communication. It also makes it hard to justify using AI agents on public-facing channels at all. Are you currently using a workaround to solve this problem? (If yes, please explain) (1-2 sentences)No, we haven’t found any workaround, since the footer isn’t part of our instructions or procedures and can’t be edited. Our only alternative would be to stop using the AI agent on this channel. What would be your ideal solution to this problem? How would it work or function? (1-2 sentences)An option to enable, disable or customize the AI footer per AI agent and per channel, with support for localized text. We’re happy to disclose AI use; we just need to control the wording and where it appears. 

Close AI Agent tickets after a short time so subsequent conversation becomes a new AI Agent ticketAccepted

We've learned that AI Agent tickets are open for four (4) days after after the last communication from the user.  This duration length is not configurable. Four days duration between the AI Agent ticket being opened and then automatically closed is arbitrary and long, isn't it. A given user might have multiple, distinct questions during a given week.  Imagine a case where a user has two questions - one on Monday and the second on Tuesday (or one in the morning and another in the afternoon).  We think it would be better for each conversation to be a distinct ticket (rather than the initial AI Agent ticket being appended to). And having the initial ticket be in the Open status for a long period of time after the conversation has concluded is a bit confusing to an agent that is reviewing AI Agent tickets. Our bot's workflow is such that a user should be able to get their answer within a few minutes (any bot should answer the user quickly, not in hours or days). So it would be better for the bot (or channel, or brand) to specify the duration rather than it being a global default. And it might also be better if the workflow could be defined such that, if a user reaches certain steps of the flow, then the flow and the ticket should be considered solved. I.e. the logic of the flow has driven the user to receive the answer they seek and there is nothing more to do with the ticket. This is so the initial ticket is quickly solved/closed and a subsequent conversation becomes a different ticket.

Sequential inactivity triggers & native Session Close- AI AGENTFeedback submitted

Please give a quick overview of your product feature request or feedback and note who in your organization is affected by this issue [ex. agents, admins, customers, etc.] (2-3 sentences)Two enhancements for the AI Agent Advanced: first, the ability to configure up to three sequential inactivity triggers with distinct re-engagement phrasing; second, the ability to trigger an automatic session closure message after 2 hours of inactivity, similar to how Inactivity triggers currently work. This affects our admins, who cannot build optimal conversational flows, and our customers, who miss out on a proactive and clear conversational experience. What problem do you see this solving? (1-2 sentences)This will solve the high drop-off rates (most of all whatsapp) by allowing more persistent and varied re-engagement attempts before a chat is abandoned. Additionally, it ensures a transparent customer experience by clearly communicating to the user when their session has been automatically closed due to time-outs. When was the last time you were affected by this lack of functionality, or specific tool? What happened? How often does this problem occur and how does this impact your business? (3-4 sentences)This is an ongoing daily issue, especially on the WhatsApp channel, where users frequently close the app, switch to other chats, or get distracted and forget about the open conversation. Currently, sending just one inactivity message is not enough to win back their attention, and because sessions close silently after 2 hours, users often try to resume the chat later expecting their previous context to be active, only to find a dead session. This lack of transparency and lost context causes significant customer friction, leading to confusion and lower bsat scores. Are you currently using a workaround to solve this problem? (If yes, please explain) (1-2 sentences)No, there is no viable workaround within the AI Agent Advanced to cycle multiple different inactivity messages or to force a final text notification exactly at the 2-hour automatic session closure. What would be your ideal solution to this problem? How would it work or function? (1-2 sentences)The ideal solution is to allow admins to stack up to 3 inactivity steps in ai agent advanced (e.g., after 15, 30, and 45 minutes) each with its own customizable text behavior. Additionally, a native "On Session Close" trigger should be added to the AI Agent settings to automatically send a final goodbye/closure message after the 2-hour timeout.