TL Consulting CEO on the right time to jump into an AI project

Stephen Marryat discusses how they help their customers take advantage of AI.

There is a growing hunger for deploying AI projects in an organisations but leaders are cognisant of finding the right time.

If leaders leave it too late, they could be left behind and have the fear of missing out but they also want to be a leader within their sector.

Stephen Marryatt, CEO at TL Consulting Group told CRN Australia about how he helps his customers to find the right time to deploy their AI project.

He said some of their smaller customers have a risk appetite for AI projects, but they are still being cautious with the deployment of them.

"We're productionising models, but we're doing it like a test bed, so they don't have as many users. They're not like an ANZ Bank that have got however many million users attached to the business model they're driving,” he said.

“If they make changes and it goes wrong or something happens in production, there's no real fallout there, so they can back it out, but it's reputational damage.

“Some of the smaller customers are kind of a different relation to risk. We're still able to progress with them.”

Speaking to their customers about AI, Marryat said many of the conversations are around governance and control more than implementation.

“That correlates to risk, there's a lot of conversations happening, but they're not materialising just yet,” he said.

"We're very strong at dev sec ops engineering and working across the Microsoft ecosystem, but I feel as though as we start to implement more models, we'll be able to become more of an influencer there, we're probably not here yet.”

TL Consulting uses supply chain solutions vendor JFrog, Marryat explained how the partner takes advantage of the company’s technology.

“When you're looking at secure supply chain, the way I look at it is holistically we've got to go in and advise,” he said.

“When we start to say how we're going to advise, you typically look at products that are common and have standardised patterns, but also are the most advanced.”

Marryat uses JFrog’s capabilities within the enterprise space.

“If you go down a little bit further, say into government, you're having different conversations at times because they may not be able to afford it, or to be able to get the expertise, and maybe a little bit of a challenge to be able to buy in,” he explained.

“It does depend, but typically speaking, JFrog is an enterprise player, so it works pretty well in that space. If you're looking at the noise, the noise is typically around security.”

In terms of agentic AI, Marryat said he is seeing more customers ask about the technology and how they can use it.

“Probably what we've done is cheekily, and this is the way I'd respond, is that we're doing it through the ecosystem play that we've created through the channel,” he said.

"If we look at say GitHub copilot, that's an example of how you would use agentic AI. Then you complement that with Databricks, where we're working with say Gen AI.

"Then with JFrog, it's ML Ops, and it's driving into the ecosystems. That's typically the way we would go. We won't create our own bespoke solutions, and there's good reason for that, because they'll be outdated pretty quickly and most businesses wouldn't come to us for that.”

Most of the time they have a “sit back and wait” approach to see how other people tackle agentic AI.

“We're waiting for that progression before people start to finish. In some cases, if I look at Microsoft, it's changing so quickly, which is great, but still not as many of those models are going to production,” he ended.