👨‍💻 Do you have customer support for your product?
What do you think about support - or have you even thought about it? =D
Do you think every product must have some kind of support?
Should it be a bot?
A human?
Email?
An AI agent?
Or all together?
In our case, with a complex workspace, I think working support is definitely a must.
We’re building it right now: a Guide Agent that helps users onboard more easily, leads them through processes, answers questions, opens feedback tickets, and more...
And there’s another interesting thing.
One of my friends works at a big company that is basically all about support - they provide support packages for other products.
Do you think this kind of business is still relevant today?
Personally, I feel like these days makers increasingly think about support as a part of the product itself, rather than something separate.
So… what do you think?
How do you approach support for your product?
Replies
I agree support should feel like part of the product especially when onboarding is involved How are you deciding what should be handled by AI vs a human
@shivam_kushwaha16Â In RunEvr, users have a support room where they meet an agent from the very first moment.
They can ask questions, talk, and get support from the agent. If they need a human response, a human support person joins the room and continues helping them =)
@adana Yeah I agree human support still matters when the issue needs context or a more personal response How do you decide when an AI agent should hand the conversation over to a human
@shivam_kushwaha16Â The user decides. The moment they ask for a human response, human support joins immediately.
@adana That makes sense I like that the user can decide when they need a human and the handoff happens right away This feels much more natural than forcing users through an AI only flow
As for a minimalist phone – yes, I am also sometimes present there :D
But for SafeLimit – not yet :)
@busmark_w_nika Is it Human Support ? or some bots or agents?
Building a smart guide agent sounds cool on paper but how does your system handle weird edge cases where the user inputs completely corrupted custom parameters?
@favour_yusuf34Â First, we suggest questions or actions for users. In most cases, they choose one and go from there.
Then, the user can simply chat and ask for whatever they need, just like in any AI chat.
Good input - useful answer. Bad or corrupted input - well, you know… =D
if your software requires an intensive onboarding bot just to explain basic navigation the core ecperience probably needs a complete redesign from scratch.
we spent months perfecting our ticketing triage workflow and the best decision we made was letting our product managers handle support rotations twice a month.
for anyone running a complex workspace, robust documentation is way more valuable than a flashy chatbot that keeps giving generic answer to specific technical questions.
We combined a fast AI tier with direct Slack access to the engineering team, and our customer satisfaction scroes jumped almost overnight withouit overloading our developers.
I'm a big Steve Jobs fan, and he shipped Apple stuff with no manual, which was crazy back then and now it's just normal. His thinking was if you need to explain how to use it, the UI's bad. I agree with that a lot. Though people really do use things differently, and there's a whole crowd of non-Apple folks out there explaining how to use Apple stuff lol Funny thing is I've automated most of my business, but support's the one place I go the other way. People want a person, and once they figure out it's a bot their satisfaction tends to drop. If a reply's going to take days though, a bot's probably better than nothing. So for me a question comes in, AI writes a draft, I check it and hit send myself, and everyone gets an answer within a day. I want to build our own chat support at some point too.
Oh and I made a RunEvr account to poke around! First thing I noticed was the chat support popping up on the right. UI and features both feel really current, looks great.
We tried setting up an automated bot during our beta phase and users hated it instantly since they just wanted a real human to troubleshoot their broken database syncs.