AI adoption stalls as NZ firms struggle to scale - Wed 9 Sep

0:00New Zealand businesses are embracing AI, but it seems many are struggling to move beyond experimentation. New research finds 98% phase barriers to scaling AI with two-thirds still stuck in the pilot phase. I spoke with concentrics as Deraj Kumar and started by asking him how hard it is for businesses to scale up AI use. Lots and lots of organisations are just simply buying technology. They think they will go by technology and that will solve the problem. Clearly, almost 9 in 10 organisations are realising that's not the case. You've got to understand why you're buying the technology, what your organisation processes are, whether they are suited to to embed these technology on top of that. How is the data fragmented today? The organisation data is everywhere. Everyone is reporting poor quality of data and not able to sort of, you know, and AI is all about treating the right sort of input, the abushing garbage out, and then the change management. So, execution is becoming the biggest challenge

1:05and coming in the way of big ambition everyone has to become an AI embedded AI led organisation. Can you put some numbers around that for us? What sort of a problem is it from your point of view that we've got here with you? Problems we are talking about 95% of the organisations are failing AI at the pilot stage. That's a pretty high hurdle rate there to get over. No wonder people are perhaps feeling a little bit like there's a party going on and they can hear the music but they're not invited if you like, that they're not necessarily seeing the immediate gains or some skepticism. How long do you think a situation like that can persist? See, the technology is evolving quite quickly and in today's world and we said this in last year, the similar report we did, you know, you've got to be, you know, the leaders in having a clear strategy organisation by and funding pathway to support the isemvation and people who understand how,

2:09you know, the technology is used and embedded into the current cell. If you don't have that, the ecosystem, you know, the higher chance of you failing and like I said, your processes, your data set, the change management, everything then comes into the play to basically making it a success or not. Are there any sectors or businesses in particular you've worked with that are doing really well that we could sort of pick up some tips from? Personally, belief banking is quite ahead. Financial services is quite ahead in one having the ambition, having a clear strategy and going after AI. However, you know, you look at, you know, pretty much all across the industry, you know, now the talks about, you know, making sure the AI you know, in fact, there's a joke about, you know, CEO going and telling the organisation, going by AI and you know, it's like, you know, we'll find a problem, you just get, you know,

3:12you just buy AI and this is just going everywhere but clearly, FS is where I see some real sort of investment, real use cases, early wins, etc, etc. I guess we've been hearing about the phenomenal growth of the companies that create these frontier models but it's more recently we've been hearing about companies that are sort of developing a use case for it and how to actually do something with that. Is that the problem that you can design this great capability but unless you've got a clear idea about how it specifically solves a problem, you kind of, people are just sort of working out for themselves badly? Oh, 100 percent, you know, you've got to be very clear about the use case and where you're embedding the AI. Lots and lots of organisations are using AI currently to sort of support humans for humans to be more efficient. That's sort of replacing human and you know, and this just keeps getting talked about in the industry that AI will come and

4:14replace humans, etc. We don't see that happening today and who knows the technology is evolving but near future as well, we see human and AI working together in bringing about that change and use cases are around, you know, when you are responding to an email query or you are on a chatbot or you are on a contact centre on a phone call with a customer is where you see the, you know, sort of embedding AI in supporting human will give you the best results and those are the organisations where they do this well in these sort of areas is where the success is and people are seeing some gains that.