Nexus Black teams go on the ground, unpack the customer’s needs and find the highest value problems that can bring the most impact using their own data, Kriti Sharma tells OGN

Organisations hesitant about artificial intelligence (AI) adoption face a clear message from industrial AI specialists: Start now, focus on hard problems, and expect measurable returns within weeks rather than months.

"There are no better problems in the world right now to solve than the hard ones in the physical environment that my team and I work on every day. We’re very focused on the industrial universe, whether it’s factories, plants, hangars, sites, fields. That’s our world," Kriti Sharma, CEO – Nexus Black, IFS, tells OGN energy magazine in an exclusive interview.

Nexus Black, IFS’s AI innovation accelerator, claims it can deliver working solutions in precisely three weeks by identifying the highest-value problems and leveraging existing models tailored to specific industrial contexts.

The firm reports clients in the utilities sector have already discovered tens of millions of dollars through optimisation projects alone, whilst the Middle East presents unique opportunities for ambitious, large-scale AI deployment from the outset rather than tentative pilot programmes.

Below are excerpts from the interview:

At the IFS Industrial X Unleashed event in New York, a few key themes were presented: ‘AI in action’, ‘Do it now’, ‘Do it fast’, and ‘do it together’. What do these mean in practical terms for organisations today when it comes to implementation of AI?

It means you’ve got to start somewhere. There’s a challenge with inaction or problem-solving. What we focus on is hard, real use cases and impact.

And I hope that kicks off a series of inspiring activities as a follow-on. Much of common imagination can be isolated to things like using ChatGPT, etc, which is what most people do.

However, in the industrial world, the stakes are much higher, and the opportunities are much bigger.

For example, in the utilities industry, it’s inspections, predictive maintenance, asset investment, planning decisions, and upscaling your workers so they can get the work done right the first time, or responding to disasters as they start to occur.

These are all hard industrial challenges, much more sophisticated than you can do with just a query and request-response system. At the Industrial X Unleashed, our focus was to raise the bar for the market.

With time and cost being a recurring theme with organisations, from a practical standpoint, can AI help organisations save both time and money?

Yes, of course. We see that all the time. AI is like any other technology, and if it’s not doing that, then what’s the point.

We see a huge amount of value, outcomes and impact already seen with our customers.

In the utilities world, for example, we’re looking at tens of millions of dollars just found in excess inventory optimisation alone.

Nexus Black claims that it can get things up and running in about three weeks. How true is that claim?

That’s right, exactly three weeks.

The audience at Industrial X Unleashed look at costs associated with the US grid on the screen

We go on the ground, unpack the customer’s needs and where the biggest value is. The chances are we’ve built something already that solves their problem. If not, we do a rapid proof of value.

What we do in those three weeks is not necessarily about making something available to every person in the company, but we find the highest value problems that can bring the most impact.

When you train the AI for a customer, does that mean training a model, like an LLM? And how long does that training take?

We don’t develop our own models as we don’t believe it’s the right thing for us to do.

It’s really about finding answers to the problems they have. That could mean using existing LLMs, tweaking them and making it relevant to their universe to enhance their existing abilities.

The training depends on the problem. But the client will see value in the three weeks that we promise, if we have access to data.

In some cases, we would create the experience and show you what the value is.

This is not point-and-click design concepts, these are real things that we have built, and we have proofs of it.

For example, it’s a challenge to build solutions that deeply understand piping and instrumentation diagrams or CAD drawings in the industrial world, and a generic AI model may not be able to do it either.

Our technology uses the same skills of an underlying model, but we make it specific to the problem that it’s looking at. That’s how we go a lot further faster.

Industrial AI is very contextual. How do you explain that?

We think AI is most powerful when it’s embedded in the workflow of the users, and it’s understanding what they do, what they’ve done before, and what they’re about to do next.

For instance, in the context of a plant or a factory, predicting what’s about to break next, puts you a few steps ahead, which means you’re driving more savings and costs rather than trying to fix something and avoiding downtime.

How do you factor in sustainability when coming up with solutions?

From a Nexus Black perspective, we help on the higher value problems and how to best apply the sustainability lens.

For instance, we track the carbon footprint at each part level through the movement of that part from A to B, and then helping identify the right decisions, like, do you make the part or do you buy it?

There’s a lot of talk about adoption of AI in the Middle East. How do you see the response of AI adoption in the Middle East?

I think it’s a unique region in many ways because the problems are unique, but also the ambition and hope, and investment, is different.

While the approach in the US or other markets is to start with pilots and grow them and scale them, in the Middle East, you see bolder action straight away; it’s about going at scale from the beginning. However, it’s thoughtful.

Nevertheless, the problems are also different. For instance, in the Western world, the use cases in utilities could be about aging infrastructure. But in the Middle East, it’s about new infrastructure altogether. In Saudi Arabia, for example, you’re putting in new transmission lines from the get-go.

We at Nexus Black only work with the most strategic clients. And for us, the Middle East is a really big opportunity and the possibility to do strategic high-value work, not small automations.

When we go to the Middle East, our conversations are about big impact, like, how do you make asset investment decisions when you’re putting in new infrastructure in Saudi Arabia? Or making hard decisions about utility optimisation, inventory decisions, maintenance at a large scale in oil and gas, or renewables?

Is a skilled workforce a challenge when it comes to the adoption of AI and learning to use those tools?

I don’t see it as a challenge, because the new wave of technology is very easy to use. It has become as simple as asking questions and getting responses, and increasing the level of automation.

I think the bigger issue is more about finding the right problems. And the limitations are identifying and knowing and being ambitious enough about which problems and use cases to go after.

What message do you have for organisations that are hesitant in applying AI?

Get someone with good boots on the ground who can come here, deeply understand your business and help you identify those problems.

By Abdulaziz Khattak