A context layer that keeps company knowledge current
Zaro keeps one current, attributable account of how a business runs — as files the work itself writes to — so every AI agent and app answers from the same governed context. I'm a founding product engineer there.
Product demo by Sammy Taylor, Head of Design at Zaro.ai · zaro.ai · LinkedIn — Zaro.ai
Context
Companies have bought AI, but it can't read the business. What it needs to know is spread across warehouses, SaaS tools, documents and people's heads — and every new AI project pays again to find that data, clean it, agree what it means and get it past governance. Zaro came out of stealth in June 2026, founded by Michael Bajwa and Qian Zheng, who had built agents at Convergence and then at Salesforce after the acquisition.
What Zaro does
Company context lives as readable files rather than rows behind an API — versioned, permissioned and attributed. The work writes back to those files, so the record stays current; any past state can be read and restored; and every answer names the file and the version it came from. Connectors bring in the systems a company already runs on, and apps and agents are built on that one shared layer — so the groundwork is done once and every later project gets cheaper.
Governed by design
AI only sees what it is cleared for. Access is scoped per file, tool and action; an agent's actions are allowed, blocked or held for review, and every one leaves a trail. Customer data is not used to train anyone's model and stays portable.
My role
I joined Zaro in 2025, and it is my main focus today. I'm its founding product engineer and designer: I build the product as well as design it, and I designed and built its mobile app, live on the App Store since September 2026.
Key outcomes
- Out of stealth in June 2026 with a $5.1M pre-seed led by Cherry Ventures, with angels including Thomas Wolf (Hugging Face) and Thomas Dohmke (GitHub)
- #3 Product of the Day on Product Hunt at launch (25 June 2026)
In the press
- The AI Insider · Zaro.ai Secures $5.1M to Unify Enterprise AI Agents, Data, and Applications in a Single Context Layer
- The Next Web · Engineers who helped build Salesforce's Agentforce raised $5.1M to build its opposite
- Tech.eu · Zaro lands $5.1M to build the next layer of enterprise AI
- Cherry Ventures · Cherry Leads $5.1M Investment in Zaro, the Context Layer for AI Agents
On LinkedIn
47 public posts- The Players Fund
“It gives us great pleasure at The Players Fund to share that Zaro.ai has joined the Juventus Forward squad.”
InvestorZaro builds a Media Desk for Juventus's press office(opens LinkedIn) - Xiaowei Zhu
“Nice interface, smooth integration with most of the popular enterprise applications on the market”
Co-host, the AWS hackathonAn outside view of the product(opens LinkedIn) - Zaro.aiMy work
“Your workspace no longer stops at your desk. Zaro is now live on the App Store, available to all existing users.”
Company pageThe mobile app goes live on the App Store(opens LinkedIn) - Michael Bajwa
“In three years, every company will have a context layer the same way every company today has a database.”
Co-founder & CEO, ZaroThe launch post — the context-layer thesis(opens LinkedIn)
All public posts (47) — team, investors, press
Zaro · 12
- Zaro.ai — What the AWS hackathon showed
- Zaro.ai — A stand at the Responsible AI Summit
- Zaro.ai — The mobile app goes live on the App Store · my work
- Zaro.ai — Invite: the BizOps hackathon at AWS London
- Zaro.ai — Guardrails Engine: per-agent permissions over the context layer
- Zaro.ai — The "Q2 FINAL v3" problem
- Zaro.ai — A workspace that remembers
- Zaro.ai — Product Hunt launch (finished #3 Product of the Day)
- Zaro.ai — What "agentic software" means
- Zaro.ai — The founding story, on the Scaling Europe podcast
- Zaro.ai — Co-founders on stage at the Cherry Ventures Summit 2026
- Zaro.ai — Launch day: the CEO reveals Zaro
Founders & team · 24
- Michael Bajwa — Zaro joins the Juventus Forward Squad 26/27
- Qian Zheng — Juventus, twenty years after staying up for their matches
- William Ross — Deploying Zaro at Juventus
- Michael Bajwa — Data teams now own the AI strategy (doesn't name Zaro)
- Qian Zheng — Measured: context reuse halves time and cost
- Nikita Pokryshko — My reshare of Qian's token-savings numbers · my work
- Qian Zheng — RAG was a workaround, not an architecture (doesn't name Zaro)
- Michael Bajwa — A busy September: dinner, hackathon, summit
- Qian Zheng — Hackathon recap: the constraint is the data
- Michael Bajwa — The first AI Transformation Table dinner
- Qian Zheng — 17 enterprise AI leaders, one seat per company
- Qian Zheng — Six months in: what Zaro is for (team photo)
- Michael Bajwa — Announcing the first hackathon with AWS
- Michael Bajwa — The biggest update since launch
- Qian Zheng — Building Zaro in King's Cross
- Qian Zheng — Teams aren't blocked on the model
- Michael Bajwa — Customer video: Scaling Europe rebuilt on Zaro
- Michael Bajwa — Five hundred licences, almost no one uses them (article)
- Michael Bajwa — Stitch your tools together, don't rip them out (article)
- Michael Bajwa — Article: a legal firm's one-line summary of Zaro
- Michael Bajwa — Why enterprise AI ends at the demo (article)
- Michael Bajwa — Article: Enterprise AI Ends at the Demo
- Michael Bajwa — Model routing is a building block; the context layer is the destination
- Michael Bajwa — The launch post — the context-layer thesis
Investors & partners · 7
- The Players Fund — Zaro builds a Media Desk for Juventus's press office
- Seb Johnson — A customer on Zaro, months after launch
- Xiaowei Zhu — An outside view of the product
- Responsible AI Summit — Zaro joins as a sponsor
- Philippe M. — An investor's thesis: AI in financial services needs a shared context layer
- Seb Johnson — Interview clip: why Cherry Ventures led the round
- Yvonne Bajela — Out of stealth: the investors and angels, named
