Ask HN: How do you think of B2B startup ideas in era of LLMs?

1 oOOmy 1 5/2/2025, 3:20:33 AM
LLMs have got darn good with writing code. Vibe coding (stupid term imo) or not, most people are getting assistance for writing code. I concur that writing code is not the most important thing in B2B but it surely makes hell of a difference for technical product.

For example, take a company that does contract management software. Common functionality is upload document, eSign, validate signature, and chain of record, integration with other tools where documents might be hosted.

When I reach out to people to validate ideas like this, I basically get no people to respond. They already have a solution or they are not interested. Michael Siebel talks about finding people whose hair are on fire and people who are early adopters for startups. I dont have network such as YC or others.

How do you validate ideas and get started for situation like this? Where do you find people to talk to that are early adopters and might be willing to give you feedback or their money?

Curious to know how has your outlook has changed with advent of LLMs?

Comments (1)

_viroo_ · 6h ago
hey, I joined a B2B startup as an early engineer and 4 years into it, found some insight in how to reach people to buy your product. First of all, the growth is slow in comparison to consumer products, and it is hard to onboard people for your product. The good thing is, once they are onboard attrition rate is low. People will incorporate your product in their process and will find it hard to change it, given you have good customer support and the product functions well. Cold emailing and reaching out to a lot of people is the only thing which will help. You can differentiate on price and try to have a simple and clean UI. B2B products should primarily solve problems and should take less time to understand and configure. Major players tend to have very complex UI, and then they require entire teams to just run that product. Try to have differentiation in these areas.