Navigating AI: EY says “ecosystem and people-first is key” to companies embracing generative AI

– Q –

WPCG: Hello Nicola, thanks for joining us. When it comes to generative AI, where do you see the biggest impacts in the near term–over the next year or so?

Nicola Morini-Bianzino: Wherever you see a human, there is a use case. This technology will affect us profoundly not only in our working lives, but also in our lives outside of work. When you think about the enterprise, you can pick an area or a function, and there is an opportunity there. I would never, ever in my wildest dreams have expected to see, for example, manufacturing companies and semiconductor companies using generative AI in R&D. We’re seeing the full spectrum of human activity effectively impacted.

– Q –

Has the technology advanced more quickly than you expected?

That’s part of it. And the second part is that the technology can do a lot of things that we didn’t expect it could do. For example, a semiconductor company here in Silicon Valley is using it to evaluate the opportunity to proceed with investments in R&D. You wouldn’t have expected something like that.

This is not about processing invoices. This is about fundamentally rethinking how you run a scientific organization. And that to me is the highlight. But then if you start going down the list, it’s everywhere. From lawyers, from tax professionals, I mean, for us, if you think about where we are as an organization, this is going to effectively turn upside down our commission model.

– Q –

If you’re a CEO, you may be feeling pressure to embrace generative AI. But the noise can be overwhelming. For a typical company that is not in the heart of Silicon Valley, what are some good ways to get started with this technology?

Regardless of what kind of enterprise you are, whether in Silicon Valley, the Midwest, Switzerland, Germany or wherever, this is about teamwork. This is not something that an enterprise, regardless of the sophistication of the enterprise itself, can do alone. For example, even if you are a software technology company, you still need the hardware providers to support you.

We leverage our technology partners across the full stack, and also clients because the interaction with clients is no longer: I’ve got a solution that you go implement. It’s more like, together we come up with new ideas and new solutions.

But the main advice is that you cannot stand still. There shouldn’t be too much anxiety yet, because what is missing in this ecosystem is the technology that ties it all together. There is a fantastic opportunity for fantastic capabilities, but a comprehensive way of connecting the dots and making sure that you can deploy at scale–that doesn’t exist yet. So you’re not going to miss the boat.

But you need to get going, and the sooner the better.

– Q –

What are the benefits of using AI at scale, across the enterprise? Is this something like the network effect, where the more people that use a system, the more valuable it becomes?

I think the network effect is the perfect way to describe it. The more people you have playing with it, the more people that can come up with different ideas. And the more they start interacting and start using it every day, the more valuable it becomes. It’s one of those things that maybe you start here, and then you end up there. It doesn’t mean that it’s the wrong thing to do. It’s just that you have to learn and use it and discover it yourself.

– Q –

When an organization is planning to scale AI throughout the enterprise, what factors must they consider? And are there tools that can help them with this?

One thing that has gone very well with our clients is the EY.ai Confidence Index. It is a methodology on how you assess the risk of a solution. Is it rocket science? Absolutely not, but it’s a great guide on how to think through a solution. For example, do you have the right talent? Do you have the hardware, because you don’t want to get halfway through training a really big model and run out of hardware capabilities. Do you have the right software platform? Do you have the right professional services companies supporting you? And what about cybersecurity? What about data? So think about it as a sort of macro checklist that allows you to assess the risk that is intrinsic to the specific solution.

– Q –

What types of flags might be raised through this process?

It could say, you know what, we’re moving sensitive data around, say, biometrics, which is a no-no. So then, in that case, the confidence is low. Not because you don’t have a good tool, but because you’re using data that is super sensitive. At that point, maybe you pull out and say let’s move to another domain. That’s the approach. It forces people to think. And I think it’s been very, very productive.

– Q –

Generative AI holds great potential, but there’s also some risks. It can produce hallucinations, where the results are inaccurate, or there’s bias within the results, or the results may contain some information that might be copyrighted or illegal. How can companies establish safeguards that ensure that their AI is operating in ways that are ethical, responsible and bias free?

You need to have very strong and robust risk management, methodology and frameworks. And you need to be looking at the problem from all the different dimensions and angles so you don’t leave anything out. Right now, you have a planet where the AI policymaking is going in completely different directions. If you are in the US, you have quite a bit of leeway on what you can do. The European Union is much more restrictive. And then other countries take a different approach. Japan is different than Singapore, and Singapore is different from India, and India is different from the UK.

– Q –

Okay, Nicola, really appreciate your time. Any final thoughts? Anything we didn’t cover that you think is important to mention?

In order not to do something that we will regret as business leaders, we need to think this technology through from all the different angles and not let the pressure and the anxiety build on us as leaders because that usually leads to bad decisions. So having a more balanced approach and using a structured logical way of doing things is probably the way to go.

Click to learn more about how AI will impact our future.