Replit Announces Acquisition of AI Business Analytics Company Atta, Integrating Its Specialized Charting and Analytics Capabilities into Replit Conversations
Replit announced the acquisition of AI business analytics company Atta, integrating its specialized charting and analytics capabilities into Replit conversations. Users can directly generate interactive and shareable charts in chat without writing SQL or repeatedly recreating visualizations.
Atta was founded in 2024 by Omar Shaik and Amine Ben Khalifa, offering a multi-agent analysis workspace on an infinite canvas: after connecting to data sources like Salesforce, Stripe, and Snowflake, the system autonomously selects analysis methods, models, and charts, producing interactive visualizations aligned with brand standards, with labels and layouts automatically adjusting to size. Omar was an early engineer for Apple Vision Pro and worked on Dropbox iOS; Amine has a PhD in machine learning and previously worked on Face ID at Adobe and Apple. About two years ago, the two cold emailed Amjad Masad for funding, and Amjad and Michele Catasta wrote a check after seeing a canvas demo that same day; subsequently, Amjad invited the team to work at Replit's office, where they were when the acquisition was announced.
Omar stated that this year, two publicly traded companies used Atta for their first quarter business review meetings; one of them replaced its existing SaaS tools with Atta documents and slides directly connected to its data warehouse. Menlo Ventures has supported Atta with SAFE since its inception and has assisted in finding design partners and the first annual contracts. Replit officially stated that investigations can turn into routine processes, conclusions can become management review slides, and then drive the next round of products or internal tools. The initial capability is live today: saying "visualize this" in a conversation will yield an explorable chart.
This marks another product acquisition for Replit following its 2025 acquisition of the integration platform OpenInt, which turned its 160+ connectors into Replit Connectors. Replit was founded in 2016 by Amjad Masad and others, valued at $3 billion in September 2025, and raised $400 million in Series D led by Georgian in March 2026, with a valuation of $9 billion. The goal is to achieve approximately $1 billion in annual recurring revenue by the end of 2026, claiming that about 85% of Fortune 500 companies already have employees using the platform. The platform has integrated Databricks Lakehouse and Microsoft Fabric, deploying generated applications into governed enterprise data layers.
Amjad framed the deal within the narrative of "autonomous driving companies": enabling non-technical business leaders to understand operational data. Funding and attention are shifting from independent BI subscriptions, data analyst queues, and slide outsourcing to Replit subscriptions and agent usage; Atta's independent brand will be integrated as a layer in conversations. Beneficiaries include teams that have built internal tools on Replit and need to bring data warehouse conclusions directly into reviews, as well as Replit's path to increasing customer spending; the pressure point is traditional BI and "chart-generating chatbots" that only sell dashboards and cannot embed building cycles.
Source: Public Information
ABAB AI Insight
Amjad Masad has been working since 2009 to lower the barrier to writing code with Codecademy, and in 2016 he turned the browser terminal into Replit, going through years of revenue platform periods and layoffs until agents expanded the definition of "code writers" to "people who can talk." Omar self-taught in a low-income immigrant community in Chicago, wrote his first line of code on Codecademy in 2012, and later worked in product engineering at Apple and Dropbox, applying the same "de-barrier" logic from programming to data reading. The first version of Atta's canvas prototype was written on Replit, and the investment and co-working occurred before the product's independent financing, making the acquisition more like formally integrating already embedded capabilities.
The capital path is for the $9 billion valued platform to use small-scale product acquisitions to fill the gap of "how to explain business after completing software." Replit has already acquired OpenInt to connect agents to external systems and now acquires Atta to enable agents to turn warehouse data into meeting-ready visuals; simultaneously, it connects with Databricks and Microsoft Fabric to govern data. The motivation is not to create another Looker, but to keep analysis within the conversation-building process, turning investigations into routines, routines into slides, and slides into the next version of internal tools, locking seat fees, usage fees, and analytical outputs into the same account.
The competitive landscape is clear: Tableau/Looker turns analysis into a standalone destination; Hex and Mode sell notebooks to analysts; ThoughtSpot sells search to businesses; Cursor and Cognition elevate coding agents to higher valuation tiers. Replit is in a phase of expanding from "generating applications" to "generating and explaining operations," similar to Salesforce integrating Tableau into CRM and Microsoft embedding Power BI into Office: channeling vertical tools.
Structural changes belong to the reconstruction of the industry chain. Reconstruction occurs where "query-analysis-presentation-decision-rebuild" is broken down into four segments charged separately: data warehouse, BI, slides, and low-code. Whoever controls the conversation and deployment environment can integrate the latter three segments into the same workflow, shifting pricing power from viewer seat charges to charges based on building and agent operation. Independent analytics startups that cannot become the default layer of a certain building platform will stall between design partners and annual contracts, waiting for the next check or acquisition agreement.
ABAB News · Cognitive Laws
Write the check first and then co-work; acquisition is just a formality.
If analysis does not enter the workflow, it is just another slide.
People who can talk replace those who can query; seats move accordingly.