Levelpath uses AI to automate the 'ugh' work that dogs enterprise procurement

Founded around the same time that ChatGPT and generative AI burst onto the scene, Levelpath offers an AI-native procurement platform that's gaining ground in the enterprise market. This summer the company raised a $55+ million Series B funding round to further its product development and growth momentum, and now has more than fifty customers. It's getting easier to pitch the AI-native solution as corporate awareness of AI's potential grows, says Alex Yakubovich, its co-Founder and CEO. He tells me:

A year ago, we would be like, 'We're an AI-native procurement company,' and we would go into these large enterprises, and they were like, 'That's very interesting, but we're not comfortable with that.' But this year, they're seeing our platform, and we go in with the AI-native procurement company message, and companies are like, 'That's what we need, because we're looking to take the stuff that used to take hours. we need it to take seconds, so that we can be way more efficient in how we're driving the business forward, because everything's moving faster.'

The young company is going up against established vendors that also claim to be harnessing the power of AI to automate procurement processes. But Yakubovich says Levelpath's all-new architecture gives it an advantage with buyers, who are wise to 'AI-washing' of incumbent products:

The great thing about selling to procurement is, procurement is like, 'Thank you salesperson for the spiel, but let's see it.' And when you show the product, the things that our customers are looking for, and prospects are looking for, is they're looking to unify. They're looking at a unified experience. They're looking at unified data. They're looking for their supplier records to be in one system of record.

Everybody's had this North Star for many, many years. We give them an application that has all of that and allows them to also have it all on an app. We call it 'procurement in your pocket.' We have native iPhone and Android apps. It's just very different. It's a very different product than you would see, particularly with any of the legacy vendors, or really anybody who was born before 2022...

We were very fortunate to be born at a time when it was so clear to build everything in a new way and build it truly AI-native, and that's what we've done. There's no part of Levelpath that is not an AI-first interface.

Supplier system of record

While the company is young, co-founders Yakubovich and Stan Garber, who is President, bring a wealth of experience as the co-founders of a previous venture, Scout RFP, a sourcing vendor that was acquired by Workday in early 2020. This background fuels their conviction that a supplier system of record — in the same way that enterprises already have systems of record for money, people and customers — should be the foundation for an enterprise procurement system. This provides the crucial system of knowledge that ensures the AI's answers and actions are grounded in the correct enterprise context. Yakubovich explains:

As AI becomes more and more capable, the really important thing — because we all have access to all the same LLMs, right? Everybody has access to ChatGPT and Gemini and all that stuff. And so the question is, how good is your context? How good is your data? And then, what are the capabilities around that?

The fact that we are built on the underpinning of supplier system of record, and that's the foundation of Levelpath, allows us to pass context into pretty much the whole platform, all the workflows, that you just wouldn't normally have. I could give lots of examples, but I'll just say that, data is more and more important every day, especially when it comes to AI. You want great answers? You need to have great context.

Equally important is a user experience that takes away all the drudgery and effort that's traditionally accompanied procurement processes. He comments:

Here's the litmus test that we look for as we build products. Can we take things that used to take hours or days, and collapse it to something that takes seconds? If you look at procurement workflows end-to-end, there's an abundance of things that take a lot of time out of procurement's day that is just done manually. Internally, we call it 'Ugh' work... This is stuff that has been done just, like painfully, over the years — through blood, sweat and tears — and we just take it away and make it really easy.

He rattles through some examples, from removing the need to read procurement manuals, to automatic collation and summarization of the documents for a procurement requisition, to instant analysis all of the terms in a contract, adding metadata, and surfacing the exceptions and risks, through automatically surfacing issues and risks when a renewal is coming up or a QBR is due. He sums up:

Think of all the time that people spend, just hunting this information down, instantly given to you as part of the workflow, as part of a report... I mean, that's why people don't run QBRs — it just takes too long to assemble the data. We have the data, we instantly cue up the workflow, get it over to the stakeholders to say, 'What's your feedback on the supplier?' And here's the full QBR data sheet on that supplier as well, in your format.

Similarly with RFPs, the system automatically generates the RFP and does the supplier research, surfacing alternative suppliers in the category that might not be on the buyer's radar, along with their contact information. He continues:

And then the bids come back. And you know how bids come back? There'll be like Excel spreadsheets, and somebody will have it out of their proposal system, and another one will do a PowerPoint. Doesn't matter, unstructured data.

If we bring it back, it instantly gives you a side-by-side comparison of all the suppliers and what you should consider, with all the data they've given you. And it'll give you TCO analysis, who's better at what parameters, et cetera, even if they submitted it in their wacky format. And of course, we have all the ways to do it in a normal submission for RFPs as well, but instantly analyzes bids. This stuff takes days. It does it in seconds.

'Rebuild end-to-end procurement'

The vendor is gradually building out these capabilities, launching each new feature when it's ready — last week it announced an AI-native contract discovery agent, which lets legal and procurement teams ask natural-language questions and get structured answers back from across a contract portfolio with clause-level references. Having started out with functions that help people start the buying process, its reach now extends right up to the Purchase Requisition (PR), which is then submitted into the customer's existing ERP or procure-to-pay system to raise a PO. Yakubovich sums up:

That's where we're at today. The full front-door-to-PR process in Levelpath — contracts, sourcing, supplier onboarding, supplier risk, all of that in there, and all the workflows and approvals that happen all along the way. Think about all the stuff that happens to make sure that doing the right thing before you put in a PR is really easy and straightforward. Levelpath helps customers with all of that.

The ultimate goal, however, is to "basically rebuild end-to-end procurement." I wondered if that meant one day automating everything to the point that AI agents will take over the entire buying and selling transaction. He tells me that's not in the roadmap:

We don't see robots buying from robots in the strategic sense... What we view as our job is, all the work that is just the drudgery, the 'Ugh' work of procurement — that I have to go through all of my contracts and figure out how tariffs are going to impact us, and where we have tariff risks in our contracts, that kind of work — that should be off of people's plates. Let the AI do it, let the assistant take care of that...

We want all of our customers to just let their [AI] assistant take care of that work, so that they could focus on building the relationship with the strategic vendors, having the conversations around like, 'How do we partner on this together and do better quality?' I don't think there's really a world where two agents are going to do a better job of building a relationship than two humans that are trying to achieve a common mission.

My take

Every new wave of technology opens up an opportunity for a new generation of vendor to build out a market footprint, and Levelpath is seizing that opportunity in the procurement space with both hands. This is a function that is often overwhelmed by the kind of documentation and research demands that generative AI is particularly good at handling — but only if it is fed the right contextual knowledge to be able to make sense of it and select appropriate enterprise outcomes. This is where the domain expertise of the founding team brings a key ingredient. The combination is finding favor with a growing number of customers, and with around $100 million now in the bank, Levelpath seems well placed to continue that growth.