Zip: Reimagining the possible in procurement

The biggest launch in company history.

For Zip, considering the growth journey the organisation has been on, that is quite the statement.

Since its founding in 2020 by Rujul Zaparde, Lu Cheng and Felix Meng, Zip has rapidly become one of the world’s leading AI platforms for enterprise procurement.

And now, at its Zip AI Summit in June located in Brooklyn, New York City, it announced three major launches – Zip Superagents, Zip MCP and AI Spend Automation. Designed to work alongside procurement teams, Superagents are an AI platform that takes on real work – reviewing contracts, unblocking approvals, coding invoices, and researching vendors – with every action audited and governed inside Zip. 

Zip MCP is a procurement-native implementation of the Model Context Protocol that connects Zip’s procurement intelligence directly to AI assistants like Claude and ChatGPT, enabling employees to work in their preferred AI tools without leaving Zip’s controls behind. 

Speaking exclusively to CPOstrategy at DPW New York City, Lu Cheng, Co-Founder and Chief Technology Officer at Zip, was full of enthusiasm about the announcement. “Superagents are the next generation of the agent foundation that we built on top of,” says Cheng. “They go beyond providing assistance – they plan, reason independently and take action autonomously. Instead of only offering guidance, they can execute tasks like sending emails, negotiating terms, redlining contracts and leveraging market and data intelligence. We are incredibly excited about the new capabilities that Superagents can unlock.

“The second launch is MCP, which is all about meeting users where they already are,” he adds. “One of its biggest unlocks is the multi-channel intake. Requesters can still initiate workflows directly in Zip, but they can also start requests from the tools they use every day, whether that is Teams, Slack, Claude or ChatGPT.”

Zip also recently unveiled its AI Spend Automation, a new enterprise offering that combines platform access, AI consumption credits, and support from Zip’s own forward-deployed engineers to help the most progressive organisations fully automate their intake-to-pay process. Block and UCI Health are among the first enterprise customers to deploy AI Spend Automation, with UCI Health already reporting $20 million in cost avoidance and value recapture from a single project. Cheng explains that UCI Health leveraged Zip’s benchmarking capabilities which allowed a single workflow to be replicated across hundreds of thousands of purchases.

“The ROI was massive,” Cheng reveals. “What we find with being able to combine agents and AI with workflows and orchestration is that the value is being able to ensure that AI is run consistently 100% of the time. Success stories like UCI Health are something we are excited about so being able to multiply that across all enterprise purchasing is the value that we see by bringing agents into orchestration.”

Governance is a major source of contention. Indeed, many companies don’t possess the infrastructure to safely run AI on their procurement data yet, however, some employees are not waiting around for it. Staff members across the enterprise are uploading spend data into Claude to analyse it, redlining sensitive contracts inside ChatGPT and generating internal financial analyses in personal Gemini or Copilot accounts. But, every time they do, sensitive enterprise data leaves systems where every action is controlled and audited, entering environments with no oversight, compliance controls or record of what was actually completed.

It is a key reason why many companies are turning to Zip. Notably, some of Zip’s customers include the companies actually building AI themselves – OpenAI and Anthropic – yet they choose to let Zip manage their procurement process.

“After working with hundreds of enterprises, including the world’s leading AI companies, we’ve learned that this kind of work is already happening, with or without governance,” Cheng says. “Even the companies building AI themselves want this work to be governed. They want the audit trail. They want the controls. They want the integrations to their systems of record. And they want to make sure their employees aren’t uploading sensitive financial, contract, and customer data into general-purpose AI. That’s why we built Zip’s orchestration layer the way we did over the past five years, so that AI doesn’t just operate alongside enterprise procurement, but inside it, where every action is governed and auditable.

“It’s been a privilege for us to be able to partner with OpenAI and Anthropic to bring leading AI capabilities into the Zip product itself. The frontier models are critical at what they do, but in a world where procurement is a large enterprise processing billions of dollars of spend every year, the model itself is really just one component of the solution. Ultimately, the model paired with the governance, auditability and the controls is what makes it successful.”

Looking ahead, Zip stands at the most exciting period in its history. A year on from launching a suite of over 50 purpose-built AI agents, Zip has pushed the envelope once again via the introduction of Superagents, Zip MCP as well as AI Spend Automation. 

“We have gone from agents that are able to help assist and provide guidance to Superagents which can act, plan, take action, reason, approve, send emails directly to vendors and can respond,” Cheng tells us. “We really think we’ve created a new baseline in terms of what’s possible. We’re doing a lot of work over the next coming months, co-developing with customers to develop best practice procurement skills, in order to enable Superagents to run effectively while continuing to invest to bring more context into our agents to make better and smarter decisions. 

“I would say the last part is continuing to invest more deeply across the entire process. When we look at going from orchestrating work to automating the downstream work, whether it’s sourcing, contract review or risk review, a large part of those components is about building out the underlying data models. We’ve natively built a lot of that functionality inside of the contract orchestration module and risk orchestration with the eye of really ensuring that we can have the right data models, projects and places to actually store the data to help automate the work itself.”

As enterprise procurement increasingly embraces automation, Zip is positioning itself at the forefront of AI-powered procurement, showcasing little sign of slowing down.

Find out more about Zip here.