Ask HN: Do you think chatbots will be like websites? | Hacker News Reader

Ask HN: Do you think chatbots will be like websites?

2 pointstomericco6 comments

Hi guys, I was wondering if chatbots are going to rise and gain revenues for companies just like websites and mobile apps did for them in the past.

Right now chatbots are efficient as websites were back in the days of Web 1.0. In most cases, they don't really help the user get things any better.

I'm talking about companies which chatbots is not their main business. Do you believe that once AI / ML will be more mature, companies will adopt chatbots as an additional communication (and revenues) channel?

Really interested to hear your take on that.

Tomer

View on HN

I think the niche for chatbots is in working with existing customers/users. Unlike a website, a chatbot is not (typically) going to attract new customers/users. Nor is a chatbot likely to provide entertainment or general information. More or less a chatbot is useful for getting 'work' done and not really for entertainment/leisure. Yeah sure there are entertaining chatbots on Slack, but the context for Slack is mostly work.

But I could be wrong.

Not really. The one way I can see it working is when we get AGI. No one would want to interact with a chatbot unless it was something they feel like they could trust.

To trust a chat bot it would need to provide real, consistent value over a long time period.

Until they can provide a better experience than current channels I don't see how it could make any kind of impact.

The experience is poor now because of the technology. AGI seems far fetched right now, so the revolution will happen only in the far future, as you say. Maybe there is no business need in these chatbots (that I described) in the first place?

Yeah I agree. Unless you can fundamentally change the value people get from this it's not worth it.

I've seen some medical based apps that have gotten funding but it's just a decision tree that tells you what you might have.

Voice will change the chatbot game ...voice recognition is already good enough...the true bottleneck is lack of methods to create generative models..at the moment all cbots run on a retrieval model.. it's likely to change in next 5-6 years..

Reply on news.ycombinator.com