Q&A: Zip, Co-founder, Lu Cheng on Contract Orchestration

Zip Forward London brought together more than 300 procurement and finance leaders. Over the space of one day, leaders had the opportunity to strategise and develop a bold new vision for the future of the industry. 

At the event, Procurement Magazine spoke to Lu Cheng, Co-Founder & CTO at Zip to hear about investments in AI adoption across procurement.

Tell us about Zip Europe and what you've launched today.

We're here in London at Zip Board Europe, and we've had some really exciting launches over the past day. We launched our new AI contract orchestration module, and we also launched several new agents for our procure-to-pay suite as well.

Let's talk about that contract orchestration — a huge launch. How does that fit into Zip's broader vision across intake, contracts and procure-to-pay?

When we started Zip, we really started by focusing on creating the intake and procurement orchestration category. A core part of our focus for the first six years was centred on orchestration, sitting on top of existing systems and existing investments, bringing together a fragmented landscape across P2P, CLM and TPRM. The average enterprise we integrate with uses more than seven systems, so we've really focused on orchestration and connecting well across those.

One thing that we've seen, especially over recent years, is that when we do a great job of orchestration, there's a natural demand from our customers for Zip to expand and take over some of the downstream processes that we integrate with. When we started, it was initially with procure-to-pay, a very natural downstream process from intake and orchestration, bringing the same ease of use from orchestration down to P2P in a single integrated solution.

And today, with AI contract orchestration, we're taking the same approach. If you look at the average legal review, it's more complex than it's ever been. Buy-side contracting is harder than ever before because the average legal review spans so many different teams. The business cares about SLAs and getting what they need quickly. Finance and procurement care about securing the right commercial terms. Legal and privacy teams care about data privacy, indemnity, and liability. And you have security teams involved who want to ensure the contracts contain the right terms to mitigate risk.

Legal review is so complex because it's deeply cross-functional. With AI contract orchestration, we're bringing the same collaborative approach from orchestration and layering in AI, not only to assist the legal review, but to actually do things like automatically redline documents and truly end-to-end automate buy-side contracting.

When we talk about AI, how do you see it becoming critical across the entire procurement landscape, or do you think it already is?

AI is absolutely critical across the entire procurement landscape. We launched our AI agents product about ten months ago, and over that period, we've deployed more than a thousand agents across several hundred of our customers. 

It's been incredible to see some of the results we've been able to deliver in partnership with those customers, spanning the entire end-to-end procurement process.

That includes helping to automate aspects of legal review, things like assisting the review of MSAs and DPAs, and working with security teams to assist the review of SOC 2 documents and, in some cases, even the full security review. 

On the procurement side, we're not only doing data validation but bringing AI to price benchmarking and adding intelligence throughout the entire procurement process.

At the same time, we've invested heavily in bringing AI to the downstream P2P process. Today, we launched agents focused on contract compliance. 

All of the hard work that goes into negotiating a contract upfront, we now have a contract compliance agent that ensures that when an invoice comes in, it complies with the negotiated terms within the contract itself.

That's been a massive unlock for some of our customers. When reviewing an invoice, no one is ever going to read through a 20- or 30-page contract to determine whether the invoice terms and line items are compliant, but our AI agent does all of that within just a few seconds. 

And with AI invoice coding, we've made significant strides beyond what traditional OCR has been able to do in P2P, really driving full automation of much of the invoice coding that happens internally.

You mentioned the results your customers are getting. What are some of the headline numbers you can share?

I can share a few examples. Some of the most impactful agents we've seen deployed have been around legal review. One of our larger enterprise customers, a very AI-forward organisation, reviews around 10,000 to 15,000 contracts a year. We partnered with them to build a legal review agent inside Zip: a legal triage agent.

The first thing it does is classify each incoming contract as high risk, medium risk, or low risk. If it's low risk, AI takes a full pass and skips legal review altogether. 

Procurement reviews the contract, AI makes the initial changes, and ultimately that client has been able to cut the number of contracts reviewed by legal by half. Around 5,000 contracts a year are now fully reviewed by AI, completely automated — and that's given them back approximately 75,000 business days of cycle time.

It's been a dramatic transformation. Not only does it help automate parts of the team's workload and remove a lot of the low-risk, more tedious steps, but it's also enabled the business to move much faster. As I mentioned, we now have over a thousand agents deployed across several hundred clients, and it's been incredible to see the results across procurement, sourcing, legal and risk.

Where do you see procurement going in this AI-driven landscape?

That's a great question. One of the things we talk about internally is that, over the next five years, we expect to see around a billion reviews inside Zip every single year, and our goal is to leverage AI to automate or assist 90% of those reviews.

When we look at what the future really looks like, many types of reviews are policy-driven and deterministic. In legal, security and risk, a lot of what happens involves taking information from a vendor and validating it internally against internal playbooks. The low-risk and medium-risk, policy-driven work, that will be fully automated by AI.

As for how the procurement role itself evolves, it will become far more strategic. It will focus not only on orchestrating processes, but on building the underlying agents needed to automate the routine work. And beyond that, procurement will partner more closely with the business, focusing on higher-value items that genuinely move the business strategically and ensuring the company is getting a high ROI on its spend.