Zip AI Summit 2026 recap: Governed AI that gets to work

On June 2 at the beautiful Ace Hotel in Brooklyn, NY, we welcomed procurement, finance, and AI leaders to Zip AI Summit 2026. The skies were clear and the coffee was strong, and the agenda moved fast: keynotes into demos, demos into customer panels, panels into workshops.

That accelerated pace was indicative of the future Zip introduced at the Summit.

We set out to spend a day showcasing what AI is capable of, what it takes to deploy responsibly, and where the technology is taking us next. Our incredible customers in the room did much of that work for us.

And we had some news to share, based around the argument that “intelligence” alone is insufficient for enterprise procurement. AI requires governance, context, auditability, and orchestration to do lasting and scalable work inside financial processes. Everything we announced is built on that premise.

Zip Superagents: AI agents that do real procurement work

The star of the show was the unveiling of Zip Superagents, a governed AI workforce that takes on real multi-stage procurement work inside Zip.

Superagents are the culmination of the Zip AI platform we’ve been building since 2023. The task-specific agents we built last year now operate as skills that Superagents assemble and coordinate.

Where a task agent runs a fixed script and produces a single output, a Superagent can reason through a multi-step scenario, hold a back-and-forth conversation, coordinate across systems, and take action the way a human user would.

Attendees saw these workflows throughout the day.

The Intake Superagent guiding an employee through a request in natural language, answering policy questions and gathering what approvers need. The Contract Superagent reviewing an agreement, flagging risks, applying playbooks, and recommending redlines. The Procurement Superagent unblocking stalled requests and managing supplier communications. The AP Superagent coding invoices and handling exceptions.

Every action runs under the same roles, permissions, and policies that govern your employees, with complete audit trails, observability dashboards, and human approval where it counts.

When your auditors ask how a decision was made six months ago, the Superagents provide the answer.

Zip MCP: governed AI inside ChatGPT, Claude, and Gemini

Something we hear from customers constantly is that their teams already use ChatGPT, Claude, Gemini, and other AI tools every day. They often research suppliers, review contracts, and draft communications in these tools—however these tools rarely operate with governed guardrails. Too often there is no audit trail, no policy enforcement, and no consistency across the team.‍

Zip MCP is the answer. It's a procurement-native implementation of the Model Context Protocol, the emerging standard for connecting AI systems to enterprise applications and data. 

Through Zip MCP, the AI assistants your employees already prefer can access procurement context and kick off procurement activities, while Zip remains the governed execution layer underneath. Permissions, controls, workflows, and audit requirements all stay intact, no matter which assistant sits on top.

Zip MCP makes these tools safe for procurement work. Whatever models and assistants win in the years ahead, Zip is built to be the layer that governs what they do with your spend.

Zip AI Spend Automation: from buying AI to deploying it

The third announcement was about execution. Buying AI technology and successfully deploying it are two very different challenges, and the gap between them is where enterprise AI initiatives often stall.

Zip AI Spend Automation combines platform access, AI consumption credits, and dedicated forward-deployed engineers who embed with your team. 

FDEs identify the highest-impact use cases, design and build Superagents and custom applications around your specific workflows and policies, define success metrics before deployment, and keep tuning as your business evolves.

Most customers will run far on out-of-the-box Superagents configured with their company context. For enterprises that want deeper customization and end-to-end automation of intake-to-pay, this is the path.

AI in procurement insights from OpenAI, Block, Datadog, and Barings

The product news was the framing for the day of forward-thinking, but the customer sessions carried the message of what’s truly possible.

In a much anticipated fireside chat, Mithun Sharma, CPO of Block, described how his team is building a culture where employees become builders themselves, now with 38 agents in deployment. At a company at the forefront of AI transformation, Mithun is driving the ‘builder’ mindset.

Later in a panel discussion, Kat Devlin of OpenAI walked us through procurement transformation at the frontier lab, including how their AI-supported workflows cut request validation time dramatically. Michelle Vita from Datadog then shared how her team identifies highly manual workflows as the first candidates for automation, while Lesley Garofola from Barings dug into supplier processes and governance. 

A theme kept resurfacing across these conversations: successful AI deployment depends as much on data quality, process discipline, and operating models as on model performance. 

Our AI readiness workshop made that concrete, giving attendees a structured framework to assess their own data, workflows, and governance against what scaled AI requires. 

The gap between organizations preparing for AI and those still working through fragmented environments is real, and it's widening.

Superagents hit the road at DPW New York

AI Summit wasn't the only place Superagents showed up in NYC. Our team was also at DPW New York, showcasing Superagents at the booth and hosting conversations all week, from executive remarks to customer-led sessions on how AI is reshaping procurement.

The gorgeous Zip booth became one of the top hangouts on the show floor. Between the matcha bar, the photo booth, and a steady stream of demos, hundreds of attendees stopped to pause, recharge, and talk through what's next for their teams.

If AI Summit was where we made the case for governed AI, DPW was where we got to pressure-test it in hallway conversations. We could tell procurement and finance leaders’ insights were sharper and more practical than a year ago, and that's exactly the direction we want to see our industry moving.

Zip Superagents and MCP availability

Zip Superagents and Zip MCP will be available later this summer and will be included with all core Zip products. Zip AI Spend Automation is available now for enterprise customers.

To everyone who joined us in Brooklyn: thank you. The questions and conversations in the hallways were as sharp as the ones on stage, and they're already shaping what we build next.

Want to see Superagents in action? You can request a demo or catch us on the AI Roadshow this summer.

‍