Luca Ferrari

I think the time where we open up ChatGPT and we tell it, okay, build me Jira as far away.

It's not months away.

It's not even a couple of years away.

AI today can't do a lot for you beyond, say, research, copywriting, and maybe some basic content production, but it'll do more and fast.

We were working with AI in 2010, so we're very early.

I'm a big believer in AI, very big.

However, even if it could build Jira today, it's not that easy to explain to it what you want when it's so complex.

So I wouldn't underestimate the inability of the user to get out of it what they need.

These products have been honed to customer needs for a long time.

You're already using them.

There is an investment in them in terms of data.

So not only does the tool need to get to a point where it can replicate that and with the same guarantee of performance, very difficult.

Like getting to something that works the same 95% of the time, we're super far from it.

But that's an infinitely easier challenge than something that works essentially 100% of the time.

Infinitely easier.

But also you need to be able to guide it to build it the way you want.

And as long as software is ultimately a relatively small share of wallet, if you think about it, it's not an expense people will optimize first.

It's not like a car that literally you plan your finances around.

Before that truly eats into the overall size of the market, I think we're talking a lot of stars need to align.

I think it's probably many years out.