Paradigm Reinvents Spreadsheets with Generative AI, Challenging Tech Giants

Paradigm launches to reinvent the spreadsheet with generative AI, filling in 500 cells per minute

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More The spreadsheet is up there with the document as one of the most important file types to make the transition from the physical world to the digital -- gaining lots more capabilities in the proccess. But besides Google Sheets and Microsoft Excel, it's hard for me to name a single spreadsheet provider off the top of my head. More importantly, both of those software programs were invented in the days before generative AI. But now, a new startup backed by famed San Francisco accelerator Y Combinator founded by a 22-year-old recent University of Pennsylvania graduate, Anna Monaco, is aiming to reimagine the spreadsheet for the modern white collar worker using generative AI to power each and every cell. Called Paradigm, it emerged from stealth today with a (relatively modest) $2 million seed funding round from Y Combinator, Soma Capita, and Pioneer Fund, as well as Arash Ferdowsi, cofounder of Dropbox; Harrison Chase, cofounder of LangChain; Eoghan McCabe, founder of Intercom; and for Jordan Singer, founder and CEO of Diagram, according to Fortune magazine's Sharon Goldman (formerly a senior reporter at VentureBeat!). Every cell is a gen AI enivronment Paradigm's launch promotional video showcases its capabilities, including an example scenario in which a user -- in this case, what appears to be a potential Paradigm recruiter -- creates a new spreadsheet with the company's software, then instructs it to look at Github for the most productive engineers and list them in order of their activity (most active to least active). Paradigm's software uses AI agents -- built atop of proprietary and open-source gen AI models from third-parties including OpenAI's GPT-4o and Meta's Llama family, according to Fortune -- to scour the web for the information the user desires and automatically populate the spreadsheet cells accordingly. According to Monaco's own announcement post on X, "Paradigm is 1000x faster than manual data collection, completing an average of 500 cells per minute." The video showcases this in realtime, as we see cells filling in with information after just one click by the hypothetical user. The user can then edit and augment the spreadsheet further using Paradigm's AI agents by clicking to add a new column, entering in natural language the categories of information they'd like -- in this case, combining data scoured from LinkedIn, Github, Twitter/X and more to summarize their tech stack -- and Paradigm will follow it autonomously and pull the information automatically. As Monaco's X post states, "The real power of Paradigm comes with scale: imagine having tens of thousands of interns working for you in parallel." The video goes on to show the hypothetical user adding a column that uses a database they uploaded of their own team to compare the potential engineering candidates to their current team and spot connections, as well as adding yet another column that compares each engineer's qualifications and experience to that of a job listing description and rates them out of 1 to 10 as to how well they match. Notable early users even as questions around accuracy remain It's but one example -- and doesn't really answer Paradigm's AI agents will be prevented from engaging in the notorious issues of hallucination and spotty math that large language models (LLMs) sometimes exhibit, such as miscounting the number of "r" letters in the word "strawberry" -- but it is a compelling one. Moreover, Monaco told Fortune that the company already has hundreds of early users from Google, Stanford University, Bain and McKinsey and starts at $500 monthly. It is currently accepting new users through a waitlist on its website: paradigmai.com. For enterprise decision-makers, this development represents a significant leap in productivity tools, particularly for industries reliant on heavy data manipulation, such as consulting, recruiting, and sales. The ability to automate repetitive tasks and enhance data accuracy could lead to cost savings, faster decision-making, and more efficient use of human resources. Enterprises should monitor Paradigm's progress and consider integrating such AI-powered tools to stay competitive and improve internal operations.

Exclusive: A 22-year-old is betting on 'swarms of AI agents' to radically transform the humble spreadsheet

Anna Monaco, a 22-year-old product engineer, designer, and recent University of Pennsylvania graduate, knows plenty about the woes of working in spreadsheets. "I've always wanted to automate my own busy work," she told Fortune in a recent interview. In addition, nearly everyone she knew at Penn planned to work in consulting and banking, where most of their day would be spent doing tedious, repetitive work in Microsoft Excel and Google Sheets -- researching, analyzing, and entering data into cells, rows and columns. She became convinced that using AI agents -- pieces of software that rely on large language models (LLMs) for automating tasks -- could be a game-changer. After completing internships at Google and Microsoft, along with a couple of startup ventures and a passion for AI, Monaco packed her bags and moved to San Francisco in January to participate in startup incubator Y Combinator: "I kind of convinced Penn to let me graduate early," she said. There, she developed Paradigm, a platform that enables swarms of AI agents to autonomously pull from millions of vetted web and third-party data points to automate spreadsheet tasks. Paradigm emerged from stealth today with a pre-seed funding round of $2 million backed by YC, Soma Capita, and Pioneer Fund, as well as Arash Ferdowsi, cofounder of Dropbox; Harrison Chase, cofounder of LangChain; Eoghan McCabe, founder of Intercom; and Rowan Trollope, CEO of Redis. Monaco said there has already been "tremendous interest" from VCs but that she and her two co-founders, June Lee and Michael Alfano, have decided to hold off on raising more funds until the platform is fully launched. Traditional spreadsheets have column headers that describe each row's information, but Paradigm offers column prompts to help instruct each cell's own AI agent that then searches and analyzes data from public and private databases including Google, Crunchbase, Apollo, and Hunter.io. The platform can also tackle tasks with several steps -- for example, a single cell could contain information that requires crawling multiple sites and reasoning based on the aggregated data. Monaco showed Fortune a demo in which she created a spreadsheet list of AI companies that automatically pulled in recent product updates, news announcements, and leadership changes, as well as startups that had raised a certain amount of money at different stages and within specific dates. LangChain's Chase, whose startup helps developers build applications using LLMs, said he was impressed by Paradigm's ability to run so many agents at the same time, quickly and accurately, in a spreadsheet. The familiar grid-based user experience of the spreadsheet also "feels frictionless," he said. "They abstract away all tedious parts...such that it just works." Monaco admits Paradigm, which is currently focused most heavily on marketing to consulting firms as well as workers in recruiting and sales, is entering a highly-competitive space that includes other AI agent or spreadsheet-focused startups like Hebbia, Wordware, and Equals AI. And, of course, Microsoft has its AI Copilot product in Excel, while Google Gemini can be used with Sheets. In addition to the new funding round, Paradigm is releasing a test version of its product to a limited number of users; others will go onto a waitlist. Monaco says among the hundreds of early users are individuals at business consulting firms Bain and McKinsey, along with Google and Stanford University. Paradigm starts at $500 monthly for businesses, but the company plans to target large customers with annual contracts. Monaco declined to disclose her company's current revenue. She insists Paradigm's AI-native product is differentiated by a seamless user experience and focus on data accuracy, as well as partnerships with data providers that enable the company to get proprietary data other tools can't provide. In addition, she said that Paradigm introduces a new form of macros, or repetitive actions. "Think Excel macros, but with prompts, where other cells, columns, and spreadsheets can serve as context," she explained. "This is a unique feature that no other tool currently supports." She added that LLMs like OpenAI's GPT-4o and Meta's Llama have become far cheaper to run in applications than when ChatGPT was first launched in 2022, making Paradigm an affordable tool for tackling spreadsheet data at scale: "We had a customer that generated data in 122,000 spreadsheet cells on our product," she said without identifying the customer. Growing up as a shy child in New York City, Monaco says she always had the desire to build. While she did not have access to coding classes, she taught herself to code by building games. "They were very simple, but they had design touches that made them really cool," she said, recalling a Flappy Bird-style game featuring other characters and settings. "I made one with a penguin and a bunch of icebergs, and I gamified it with a leaderboard so people would literally compete in class all day," she said. "I would look around at the computers in front of me, and half of them would be playing like the game I made." Besides building Paradigm, Monaco's busy summer of 2024 also included co-founding (and living in) an all-female hacker house in Silicon Valley (called HackHer House, located in the Russian Hill neighborhood of San Francisco) where women founders pulled all-night coding sessions and helped out each other's companies. "It's just been meaningful to have a lot of other really, smart, inspiring, like-minded women around me," she said, adding that she feels a "responsibility and a desire to allow like more women to get really into [building AI], just because I think it's great." Nowadays, Monaco and her two cofounders live and work together in a different house in San Francisco's Twin Peaks neighborhood, where, she says, the trio bring a similar "hackathon" energy to their work. "It is literally like we live in a hackathon every single day to build this sort of thing in the amount of time that we did," she said. "It is nonstop, but we actually love it." That focus, apparently, leaves no room for fear. Monaco maintains she is unfazed by the competition, including Big Tech. "I think that people get scared away from startups that are actually solving the biggest problem in the most competitive space," she said.