Zaro lands $5.1M to build the next layer of enterprise AI

Overview

Zaro, a London startup, left stealth on June 9, 2026 with $5.1 million in pre-seed funding led by Cherry Ventures. The eight-person team, made up largely of engineers who built AI agents at Convergence and helped ship Salesforce's Agentforce, is building an AI-native workspace that ties a company's tools, workflows, and data together through a single shared context layer the business owns. Angel backers include Hugging Face co-founder Thomas Wolf and GitHub CEO Thomas Dohmke.

Key Takeaways

Zaro raised $5.1 million in pre-seed funding, with the round led by early-stage firm Cherry Ventures.
The product is a shared context layer that connects a company's data, decisions, workflows, and operational history so AI agents and apps stop working in isolation.
The founding bet is that AI models grow interchangeable over time while a company's accumulated knowledge stays unique and compounds in value.
Five of the eight team members previously built AI agents at Convergence, a startup Salesforce acquired, and worked on Salesforce's Agentforce.
Angel investors include Hugging Face co-founder Thomas Wolf, GitHub CEO Thomas Dohmke, Charlie Songhurst, Mandeep Singh, and Convergence co-founders Marvin Purtorab and Andy Toulis.
Zaro routes routine tasks to cheaper AI models and reserves advanced models for complex work to lower running costs.

Stats & Key Facts

#$5.1 million raised in the pre-seed funding round
#1 lead investor, Cherry Ventures, backing the round
#8 people on the founding team
#5 team members who previously built AI agents at Convergence and worked on Agentforce
#6 named angel investors in the round, including two well-known tech-company founders

Zaro Exits Stealth With $5.1M Led by Cherry Ventures

The company went public with both its product and its funding on the same day.

Zaro came out of stealth on June 9, 2026 with $5.1 million in pre-seed funding. The round was led by Cherry Ventures, a European early-stage investor. Pre-seed money of this size is meant to fund the earliest stage of a company, before a product has many paying customers.

The London-based startup said the cash will go toward building out the product, growing the team, and getting the platform ready for larger enterprise customers. Coverage of the deal framed Zaro as an 'anti-vendor' play, a description tied to where its founders came from and what they are trying to build.

A Shared Context Layer the Company Owns, Not the Vendor

The product centers on one idea: a single place where company knowledge lives and accumulates.

›The platform connects a company's data, decisions, workflows, and operational history in one layer.
›AI agents, applications, and automated workflows run on top of that layer instead of each holding its own slice of knowledge.
›Information produced in one process feeds future tasks across the whole organization.
›Teams build custom applications using their own documents, data, and processes.
›A marketplace offers pre-configured workflows so companies do not start from scratch.

The Problem: AI Tools That Work Alone but Not Together

The founders say the core failure of enterprise AI today is fragmentation.

Most businesses bolt on AI agents and automation tools one at a time, and each tool keeps its own narrow view of the company. As a result, knowledge created in one system never reaches the next, so the organization never builds a single compounding memory.

Chief executive Michael Bajwa summed up the gap by saying the team built agents that worked flawlessly in isolation and then watched them struggle to work together. Zaro's answer is to give the whole organization one context layer it controls, rather than handing that role to individual software vendors.

Why Context Beats Models Over Time

Zaro's thesis is about where lasting value sits in enterprise AI.

Chief technology officer Qian Zheng argues that context compounds. In his view, AI models become increasingly interchangeable as the market matures, but the value an organization creates from its own accumulated knowledge stays unique to that company.

That belief shapes the whole product. If models are swappable but context is not, the smart asset to own is the context layer, not any single model. Zaro positions itself as the place where a company holds that asset for itself.

The Founding Team and Its Agentforce Roots

The team's background is central to its pitch.

›Co-founded by Michael Bajwa, chief executive, and Qian Zheng, chief technology officer.
›The team has eight people, five of whom previously built AI agents at Convergence.
›Convergence was acquired by Salesforce, and team members went on to work on Salesforce's Agentforce.
›The founders say they left because companies running Agentforce were feeding the vendor's context layer instead of their own.

Multi-Model Routing to Hold Down Costs

Cost control is built into how the platform runs AI work.

Zaro uses a multi-model approach. Routine tasks are routed to lower-cost AI models, while more advanced and more expensive models are reserved for complex workloads. The aim is to reduce operating costs compared with deployments that rely only on top-tier frontier models.

This matters to non-technical buyers because frontier-model usage at enterprise scale gets expensive fast. Sending the simple work to cheaper models is a practical way to keep AI bills predictable as usage grows.

Backers and the Startup Running on Its Own Product

The angel list and Zaro's internal use both speak to early credibility.

›Angel investors include Thomas Wolf, co-founder of Hugging Face, and Thomas Dohmke, chief executive of GitHub.
›Other angels are Charlie Songhurst, Mandeep Singh, and Convergence co-founders Marvin Purtorab and Andy Toulis.
›Zaro runs its own platform internally for functions such as HR, finance, and facilities.
›Using its own workspace lets the team test how the shared context model holds up across real business processes.

Frequently Asked Questions

What does Zaro actually sell?

Zaro sells an AI-native workspace built around a shared context layer that connects a company's data, workflows, and operational history. AI agents and applications run on top of that layer so knowledge created in one process informs work across the whole organization.

How much did Zaro raise and who led the round?

Zaro raised $5.1 million in pre-seed funding. The round was led by Cherry Ventures, with participation from several angel investors.

Who are the angel investors backing Zaro?

Named angels include Hugging Face co-founder Thomas Wolf, GitHub chief executive Thomas Dohmke, Charlie Songhurst, Mandeep Singh, and Convergence co-founders Marvin Purtorab and Andy Toulis.

Why is Zaro described as the opposite of Salesforce's Agentforce?

Several Zaro team members helped build Agentforce, then left because companies using it were feeding the vendor's context layer instead of owning their own. Zaro flips that model by giving each business a context layer it controls itself.

How does Zaro keep AI costs down?

Zaro routes routine tasks to lower-cost AI models and reserves more advanced models for complex work. This is meant to reduce operating costs versus running everything on top-tier frontier models.

Zaro is betting that the lasting value in enterprise AI sits in a company's own accumulated context, not in any single model, and its $5.1 million pre-seed round gives a team of former Agentforce engineers the runway to test that idea. Whether businesses adopt a vendor-neutral context layer over the established platforms is the question the funding now sets out to answer.