DPW New York: Inside Levelpath's Future AI Blueprint

DPW New York 2026 saw procurement leaders and innovators gather to discuss the function's future and what it takes to shift the collective dial from AI adoption to execution. 

The event, which took place on 3-4 June was centred around a 'RECODE' theme. This overarching narrative was focused on how to maximise competitive advantage over the next decade by rethinking and redesigning procurement around AI developments now and in the future. 

Those gathered agreed that now is the time for firm action, to transform operating models and strategic leadership, and to harness technology. 

It was the perfect environment for companies like Levelpath, an AI-native procurement platform designed to help global enterprises manage their purchasing, supplier relationships and contracts. 

Co-founder and President Stan Garber spoke with Procurement Magazine at the event, discussing why the company's AI-native approach matters, how it is leading the shift to agent operations and his blueprint for AI procurement that scales. .

Moving from just technology adoption to real measurable results is a huge theme. Where does LevelPath sit with that, and why does procurement need an AI front door and an orchestration platform right now?

I love that question. As technology is evolving, one of the things that won't change is the outcomes at the top. The deliverable measures – whether it's savings or more spend under management – will be concrete things for procurement's existence. 

But how the job gets done is evolving. We see a world in the next few years where you actually have components of procurement being automated.

Autonomous procurement is the dream. It's not going to happen tomorrow, but that is where things are going.

The idea of having agents actually be part of the metrics is key. One of the things we talk about at LevelPath is a new set of metrics that CPOs and procurement organisations have to adopt: the number of workflows automated, and the number of agents that are actually doing tangible things. 

When you do the scorecard and you report out to the C-suite, you’re then able to show that. Companies like LevelPath are the enablers to do that.

You’re AI-native. Why is that native approach so important from the outset?

One reason is the speed of innovation. LLMs functionally change their capabilities every six to nine months and, when you're built on or bolted onto a legacy core stack, you won't be able to keep up.

When you think about the word ‘native’, a good parallel example is your phone. There's the mobile web where you go to a website, but you can't do everything you want on it. Then there's a native app built within the app store that uses all the capabilities of the mobile device. 

It is the same idea now with an LLM. You're either built around the LLMs and using the native capabilities within the LLM, or you're bolting it on and just using a few of the core features. 

When we say AI-native, LevelPath is a next-gen product that's actually built with the LLM as the underlying technology that goes through the actual datasets all at once.

You’ve spoken about a shift from standard operations to ‘Agent Operations’. What does that actually look like in practice?

One really big thing that will change in the next few years is that every group has procurement operations – the individuals that come in and set the capabilities of how the software will get used. 

The next evolution of that role will be agent operations.

The reason I say that is because not every single person in the company should be able to build their own agents. Writing an agent does require some level of skill. 

Having those individuals codify your company's processes and then automate them ensures that every time you come in, it performs a behavior that you would want it to do – like a contract review, a sourcing analysis, or filling out an intake form from an uploaded contract or invoice. 

You want that consistency. There will be a set of individuals that build this, and then they can tweak it. That is a functional role that every CPO should be thinking about.

You recently laid out a blueprint for AI procurement that scales. What are the core steps of that playbook?

Our blueprint relies on three simple things: one, up-leveling your team. Actually getting them to experiment with AI and putting it as part of their goals and quarterly performance is step one. 

Two, agent operations – creating a functional role that actually helps build agents through that team. 

Three, new KPIs. KPIs that you can actually hold people to, because what you can't measure, you can't get results on. 

The simplest KPIs to start with are the number of workflows automated and how many agents you have running, followed by the actual results those agents have given. 

Those three things are very simple for CPOs to start doing tomorrow.

What excites you most about still being in the game, and what will really shift the dial over the next few years?

Right now, being in tech and this industry, it is the golden era of building software. 

We've never been able to move as fast as we are moving today. We've never been able to see as much impact as we have in such a short amount of time, and you can drive that impact with much smaller teams.

What gets me excited is this whole idea of autonomous procurement. It's not going to happen overnight, but you can create so much automation for the mundane things that need to happen day-to-day over the next five years. 

Procurement will look very different five years from now, and to be a part of that technology being built right now is a game-changer.