Overview

Vizly was a Y Combinator-backed startup (S23) that built an AI-powered business intelligence tool allowing users to query and visualize data using plain English.Founded in 2023 by Ali Shobeiri and Sami Sahnoune, two McGill University graduates with directly relevant experience at Apple, Microsoft, NASA JPL, Splunk, and Plotly, the company's core thesis was that non-technical employees were bottlenecked waiting for data analysts to answer routine questions — and that a lightweight, on-premises AI layer could eliminate that bottleneck without exposing sensitive data to external APIs.

Vizly raised $500K from Y Combinator and never secured follow-on funding.It operated as a two-person team in one of the most crowded AI application categories of 2023, competing against Microsoft, Google, OpenAI, and a wave of well-funded vertical challengers.

By late 2024, co-founder Ali Shobeiri had joined Perplexity, the company's domain had been acquired by an unrelated firm, and YC listed both founders as "Former Founders." Vizly shut down quietly, with no public post-mortem.

Founding Story

Ali Shobeiri and Sami Sahnoune had known each other for nearly a decade before founding Vizly. ^[1] Both graduated from McGill University and had already co-built a product together: YouTube Party, a Chrome extension that synchronized YouTube video playback across multiple users and reached over 300,000 weekly active users before being sold. ^[2] That prior collaboration gave them a working template for how to build and ship together quickly.

Their professional paths after McGill took them into the exact domain Vizly would eventually address. Shobeiri worked as a Machine Learning Engineer at Apple, a ML Research Intern at Unity, and a Software Engineer at Microsoft. ^[3] Sahnoune built mission-critical data systems at NASA's Jet Propulsion Laboratory, architected data visualization libraries at Splunk, and worked at Plotly. ^[4] These were not peripheral roles — both founders were embedded in data infrastructure at organizations where data access was a daily operational constraint.

The founding insight came directly from those experiences. At NASA JPL, colleagues in cybersecurity would regularly approach Sahnoune with questions about datasets, creating a recurring bottleneck where a single data-fluent person became the gatekeeper for an entire team's analytical work. Shobeiri experienced the same dynamic at Apple, where non-technical colleagues were queued up waiting for him to answer data questions. ^[5] The problem was not hypothetical — both founders had lived it from the analyst's side, fielding requests they believed should not have required their intervention at all.

The initial vision was to eliminate that bottleneck by letting non-technical users ask questions of their data in plain English and receive charts, summaries, and insights without writing a single line of SQL or Python. The on-premises deployment model — a deliberate architectural choice, not an afterthought — was designed to make the product viable for enterprises in regulated industries where data could not leave internal infrastructure.

The team began building before YC acceptance. Their first public prototype appeared on Hacker News in May 2023, ^[6] months before the S23 batch officially launched. That early public test reflected a bias toward shipping that would characterize the company's product cadence throughout its short life.