Share product ideas and upvotes with our product team
There’s a really straightforward bug I tried to report yesterday but had to have a painfully long “conversation” with Fin. It’s been a day now and I was prompted regarding the long wait, whether I still need assistance or not.It feels really silly to wait to speak to a person (after giving all the bug information to the bot) for information like this that can simply be relayed and read whenever Intercom staff has the time.I’d really like to see a way to bypass Fin to directly report bugs or give feedback on the product. I’d prefer if Intercom reach out to me if they don’t understand my bug report or product feedback, rather than needing to be available to reply whenever they get a chance to check my message. I think it would save both customers and Intercom staff’s time.(FWIW, the bug I want to report is that when you upload a user csv without a header row and it has errors, each line it reports an error on is off by 1 row. For example my error report said 3, 11, 17, 25 but those entries were imported correctly and it was 4, 12, 18, and 26 that had an import error. So it took me a long time to figure this pattern out because I felt the need to check every single entry, since I was having both false positives and false negatives when I spot-checked the list.)
I think it would be really helpful if, as well as qualifiers, we could choose exceptions when writing a custom answer and training Fin to pick up on the right questions.E.g. you can set exceptions when you create a Topic; words that you want the system to ignore, and if that word appears, then the topic is NOT applied to the conversation.I think this would be really useful for creating Custom Answers, as qualifiers don’t always help, but suggesting words to ignore would stop Fin giving the custom answer to similar questions where it’s not actually the correct answer.Anyone else think the same?
Is it possible to create a custom bot (Not FIN) with chat GPT integration?
Current Situation:At the moment, customers can select a category for their inquiries in the Intercom Messenger, which works well for directing questions to the appropriate teams. However, for customer inquiries submitted via email, this categorization process doesn't work, causing all emails to be assigned to a single team by default. This creates extra manual work to redirect these inquiries to the correct team.Proposed Solution:We would like to leverage AI to analyze the content of customer questions and automatically assign them to the most relevant team, regardless of the channel through which the inquiry is submitted (Messenger or email).How it Would Work: AI-Based Content Analysis: The AI would analyze the text of each customer inquiry and match it with the most relevant team based on the content of the message. Team Descriptions: For the AI to function effectively, each team within our organization would have a detailed description of the types of inquiries they handle. These descriptions would be used by the AI to make accurate assignments. Cross-Channel Implementation: This solution should apply to all channels, including both Messenger and email, ensuring that inquiries are routed efficiently without requiring customers to manually categorize their questions. Reduced Customer Effort: This removes the need for customers to select a category or specify a team, allowing them to simply ask their question. The AI takes over the categorization process, resulting in a more seamless experience.
We have hundreds of Custom Answers within our workspace, and it’s fairly challenging to search for them within Custom Answers menu, since there is no dedicated search field. At the moment we are limited to tagging and searching with Command + F for the name, but there is no way to search based on the Custom Answer content. This makes using Custom Answers unscalable for bigger teams, or forces us to use very strict naming conventions. It would be great to be able to create custom folders (and sub folders) for custom answers that could be nested under Custom Answers menu in the UI to simplify content management
We have numerous Slack escalation channels which contain a wealth of knowledge - it would be very useful to be able to give Fin copilot access to particular Slack channels to draw from.
In the future it would be really helpful if Fin AI Copilot could read and use tracker and back-office tickets to generate answers.
In addition to the settings currently available for Fin AI Copilot where we can limit the conversations by teammate that Copilot will use, it would be great to be able to mark individual conversations as incorrect or “do not use” so that Fin will not use the answers from that conversation again in the future. This would be great for when a behavior or answer changes with a release (and before release the previous answers can be marked as “do not use” to ensure that the new answer is correctly provided from new articles and other resources instead of old conversations) This would be helpful for one off or unique situations that come up where an incorrect answer was provided to a customer and this conversations like this can be marked as “do not use”
I can see that the suggested replies have been improved a little and now include a ‘Thanks for reaching out’ in addition to the greeting line. It’s always set to the line immediately underneath it though, which isn’t how I compose my messages. I always put a space between these lines, so I can never use the auto composer.Can this be updated to have a space between the lines?
I often find myself copying what I’ve put into a Custom Answer or Workflow and going back to a conversations, putting in what I’ve written and clicking “rephrase” or “make more friendly” It’s one of my favorite features and we would love to use it in every area of Intercom!
Already have an account? Login
No account yet? Create an account
Enter your E-mail address. We'll send you an e-mail with instructions to reset your password.