# Users get stuck during onboarding in our app; what in-app bot platforms can guide them and route issues to support? By Mantas Aleksiejevas > What an in-app bot needs to get stuck users past onboarding blockers, when and how it should route issues to support, and how the platforms compare. URL: /blog/in-app-bot-stuck-users-route-to-support Published: 2026-10-08 Target keyword: in-app bot stuck users route to support ## Quick answer Look for an in-app bot that does two things well. First, it guides the stuck user past the step on their own screen, using your help docs, and does the step for them where you allow it. Second, when it cannot help, it routes the issue to your support team with the context: who the user is, what they were doing, what they tried and where they got stuck. Support bots are strong at routing but wait for a question. Tours guide fixed steps but rarely route. A live onboarding agent such as Obi guides the user and hands over with the context. ## Guide first, route second A stuck user wants to keep going, not open a ticket. So the bot should try to get them past the step first. Most onboarding blockers are small: a field in the wrong format, a missing permission, a setting on a screen they cannot find. A bot that sees the screen can often clear these quickly. Routing is for what the bot cannot fix: a bug, an account problem, a custom setup, or a user who asks for a person. For why users get stuck in the first place, see our guide to [AI in-app onboarding assistants](/blog/ai-in-app-onboarding-assistant). ## What a good route to support looks like ### Clear rules for when to route Write down when the bot hands over. For example: the same step fails twice, the user asks for a person, the problem looks like a bug, or the account has a named contact on your team. ### Context your team can act on The handover should say who the user is, which step they were on, what they tried and what the bot told them. Your team should not have to ask the user to start again. ### It lands where your team works Check where the handover goes: your helpdesk, your chat inbox or Slack. A handover that sits in a separate dashboard gets missed. ### The user knows what happens next Tell the user who will reply and roughly when. Then let them carry on with any steps that do not depend on the fix. ### Every handover teaches you something Group handovers by setup step. The steps with the most handovers show you which help docs to fix and which product screens to change. ## Questions to ask vendors - **Does it see the user's screen?** Or only what the user types? - **Can it do the step for the user?** And can you limit which steps? - **Can you set the routing rules?** By problem, plan or number of tries. - **What does a handover contain?** Ask to see a real one. - **Where does it land?** Your helpdesk, your chat inbox or Slack. - **Can you see handovers by step?** Not just a list of conversations. ## How the main platform types guide and route | Platform type | Examples | How it guides stuck users | How it routes to support | | --- | --- | --- | --- | | AI support platforms | [Intercom Fin](https://fin.ai/), [Zendesk AI agents](https://www.zendesk.com/service/ai/ai-agents/) | Answers the user's question from your help center in a chat widget | Passes the conversation to your support team in the same helpdesk | | Product tours and adoption tools | [Appcues](https://www.appcues.com/), [Pendo](https://www.pendo.io/), [Userpilot](https://userpilot.com/), [Whatfix](https://whatfix.com/) | Shows fixed steps and tooltips that you build ahead of time | Usually through a link to your chat tool or help center | | Live onboarding agents | [Obi](/product/onboarding-agent) | Sees the user's screen, guides them past the step and can do it where permitted | Hands over to your team with what the user was doing and where they got stuck | Many teams pair a support platform with an onboarding agent. The agent guides setup, and the support platform handles tickets. See our reviews of [Fin](/blog/reviews/fin), [Zendesk AI agents](/blog/reviews/zendesk-ai-agents) and [Userpilot](/blog/reviews/userpilot). ## Where Cor fits Obi is Cor's AI onboarding agent. Your team can place Obi inside your product, on the screens where users get stuck. When a user stalls, Obi sees their screen, works out the cause from your help docs and guides them past it. Where you permit it, Obi completes the step after the user confirms. Your team sets when Obi hands a user to a person. The handover carries what the user was doing and where they got stuck. If Obi cannot answer a question, it says so and flags it in the session report. An issue log shows the steps where users get stuck most. Teams using Obi see up to 2.5x faster onboarding. Keep your support platform for tickets. Obi takes the stuck moments during onboarding. [See Obi inside your product](/product/lifecycle-agent) See [customer onboarding software to reduce support tickets](/blog/reduce-onboarding-support-tickets). ## How to choose - If users stall on setup steps without asking, pick a bot that sees the screen, like a live onboarding agent. - If most issues are account or billing questions, a support platform may be enough. - If you already have a support platform, make sure the onboarding bot's handovers reach the same team. - If your product is simple enough to learn by clicking through, product tours may be enough. - Many teams pair the two: an agent for setup and a support platform for tickets. ## About the author **Mantas Aleksiejevas** Co-founder of Cor. Works on Obi, the AI agent that runs customer onboarding. [LinkedIn](https://www.linkedin.com/in/mantasaleks) ## Frequently asked questions **What should an in-app bot do when a user is stuck?** Try to get them past the step first, on their own screen. If it cannot, hand them to your support team with the context and tell them what happens next. **When should the bot route an issue to support?** When a step fails more than once, the user asks for a person, the problem looks like a bug, or the account has a named contact. Set these rules up front. **What should a handover to support include?** Who the user is, which step they were on, what they tried and what the bot told them. Your team should not need to ask again. **Can an in-app bot work with our helpdesk?** Ask each vendor where handovers land. They should reach the tool your support team already uses every day. **How do we reduce the number of stuck users over time?** Group handovers and session reports by setup step. Fix the help docs and product screens behind the steps with the most handovers.