Rows AIreview, pricing and verdict

Last updated: 2026-07-01

Rows AI was an AI spreadsheet whose own benchmark claimed a 74% fully-dynamic output rate, ahead of Google Sheets and Excel, on real-world data tasks. It ran GPT formula functions and natural-language analysis inside a normal grid instead of a chat window. Superhuman acquired Rows and shut the product down in 2026.

About Rows AI

Rows AI was an AI-native spreadsheet platform built by Rows, a Berlin-based startup founded in 2016 by Humberto Ayres Pereira and Torben Schulz. The company raised $33.7M across four rounds from investors including Accel, Lakestar, and Cherry Ventures, and grew to serve business teams across marketing, sales operations, and finance. In February 2026, email productivity company Superhuman acquired Rows; the product fully shut down on May 31, 2026, with the team joining Superhuman to work on Coda. The core product combined standard spreadsheet mechanics with GPT-powered AI formula functions: =AI.EXTRACT(), =AI.CATEGORIZE(), =AI.SUMMARIZE(), and =AI.SENTIMENT() applied AI operations to any cell range at scale. The AI Analyst accepted plain-language queries typed directly into the grid, handling multi-step analyses like trend forecasting, data cleaning, and dataset merging while displaying its reasoning inline. In Rows' own AI Spreadsheet Benchmark of 53 real-world tasks, Rows scored 89% first-try accuracy, compared to Excel at 53% and Google Sheets at 57%, and was the only tool to produce fully adaptive, dynamic outputs on 74% of benchmark tasks. The platform pulled live data from 50+ sources including Stripe, HubSpot, Google Analytics, Salesforce, Shopify, Facebook Ads, and LinkedIn Ads via native integrations that refreshed on a schedule. Teams could enrich raw data with AI classification, translate text columns, extract named entities from free-text fields, and publish the results as interactive shareable dashboards, all without SQL or Python. HTTP integration tasks let users connect to any REST API from within the spreadsheet interface for custom data pipelines. Every pricing tier, from the free plan through Enterprise, included the AI Analyst and the AI formula functions; higher tiers scaled task volume rather than gating features. The platform was web-based with a progressive web app installable on iOS via Safari and on Android via Chrome. Rows built its own AI Spreadsheet Benchmark covering 53 tasks including arithmetic, data summarization, pivot construction, and column manipulation, publishing the methodology openly in early 2026. Before the shutdown, the platform integrated with Google Sheets for data import, supported OAuth connections to major SaaS platforms, and allowed API access for external systems to read and write sheet data programmatically. That API access, along with the rest of the product, ended with the May 2026 shutdown.

Pricing

Free plan: 500 AI enrichment tasks per month, no credit card required. Plus: $59 per month for 1 million tasks. Pro: $249 per month for 1 million tasks plus extra AI prompt capacity. Enterprise: custom pricing. An $8-per-seat fee applied on top of the Plus and Pro base price, so a 10-person Plus team actually cost $139 per month, not $59. The product shut down in May 2026; no new signups were accepted.

Key Features

• AI Formula Functions: Typing =AI.CATEGORIZE() or =AI.EXTRACT() into a cell applied GPT classification, extraction, summarization, or sentiment scoring across an entire column at once, turning a raw ticket export into structured data without a script.
• AI Analyst: Ask a question in plain English inside the grid, such as which campaigns are losing money this month, and the AI Analyst ran the multi-step analysis itself, showing each reasoning step rather than just a final number.
• Live Data Integrations: Native connectors to Stripe, HubSpot, Google Analytics, Shopify, and Salesforce pulled data into the grid on a scheduled refresh, so a sales or marketing team never had to copy-paste a CSV export by hand.
• Benchmark-Leading Accuracy: On Rows' own internal benchmark, the tool matched or beat Google Sheets' 57% first-try score across the tested tasks, though the benchmark itself was never independently audited by a third party.
• Automated Dashboards: Live data could be turned into an interactive, shareable dashboard with a couple of clicks, no code required, with the underlying source refreshing on its own schedule so the numbers on screen stayed current.
• HTTP API Integration: Built-in HTTP request support let a user pull from any REST API directly inside a formula, covering data sources the native connector list didn't reach, without needing a separate iPaaS tool.

Pros

• Its own internal benchmark claimed roughly 1.7x the first-try accuracy of Excel and Google Sheets combined, a large enough gap to notice, though the numbers were self-published and never independently verified.
• Native live integrations meant a marketing or ops team stopped manually exporting CSVs from Stripe or HubSpot just to refresh a report; the connection lived inside the sheet.
• The free tier included the full AI Analyst and every formula function, not a crippled trial, so a solo operator could evaluate the product without talking to sales.

Cons

• Product shut down in May 2026 after the Superhuman acquisition; no migration path to a successor product was provided.
• AI response times averaged 138 seconds per query, making interactive analysis slower than asking ChatGPT or Claude directly for the same ad-hoc question.
• Datasets exceeding 100,000 rows caused performance degradation, effectively blocking adoption by data engineers or analysts working with enterprise-scale tables.

Visit Rows AI Official Website